[PRJ][OS]Linux and Android on HTC Diamond "NOT YET WORKING"

[kal1990]

hey there, i get the same messages like ATT-Half. then it seems to hang.

i tried the .exe and the files, i turned tf3d off but nothing changed...how can i find out whats the problem?

[petay]

i have found another phone running android. well i think i have!!! the site looks a bit amateur but maybe there pooling all there efforts into the device rather than the web site!!! http://www.zzzphone.com/

maybe they could help with the port or there device may hold some drivers for us????

[kam187]

lol it looks like a 15 year old made that site for a school project

[balsat]

My programing skills is very bad, but this http://ethesys.library.ttu.edu.tw/ETD-db/ETD-search-c/getfile?URN=etd-0831107-173708&filename=etd-0831107-173708.pdf pdf file seems very usefull!?

[ATT-Half]

Did you try tapping the screen, even when it seems to hang?

Btw, it doesn't hang because of the 'rtc' message. We all get that message. The STUB messages are normal too.

RTC is real time clock (reads the clock from hardware). This also doesn't work on RAPH100. It's not causing any problems.



yes tried tapping the screen hit every possible key on the phone nothing

[Blazeitup123]

Android 1.0 gets ported to the Nokia N810


http://www.engadget.com/2008/12/05/android-1-0-gets-ported-to-the-nokia-n810/

[saphir]

Hello,
(...)
After a few hours of searching the web, i found that the Diamond is using a moviNAND flash chip from Samsung.
(...)


Could you point us to some links to confirm the use of a Samsung chip?
If you're right, porting Samsung native linux driver to the kernel should be easier than developing a new one.

[edited] The driver is not yet available on their website but http://www.linuxdevices.com/news/NS8555920711.html?kc=rss

[vomoho]

Hi,

Sorry to distrib you, for this question, I thing you have already answer to this question but i didn't the answer.

i want to know if install android on my htc (diamond) need to remove WM, because i need to keep some application like tomtom...

Thanx for your answer, and sorry for my bad English I'm French.

[itsallsubliminal]

Hi,

Sorry to distrib you, for this question, I thing you have already answer to this question but i didn't the answer.

i want to know if install android on my htc (diamond) need to remove WM, because i need to keep some application like tomtom...

Thanx for your answer, and sorry for my bad English I'm French.

android is not near full functionality yet

when you load android, it shuts out windowsmobile(completely) and boots android... by shut out, i do not mean erase... you can toy with android for some fun of having a different operating system, and when you want to return to windows mobile, just do a soft reset, thats all...

your windows mobile applications will NOT work on the android, nor will you have your contacts in android

[Unimaginative]

Not sure if this would be any help, re: getting Android to see the NAND:

"YAFFS2 -- 35 Files

Unlike PCs, which store files on disks, mobile phones store files in solid-state flash memory chips. The HTC G1 uses NAND flash, a type of flash memory that is becoming more popular due to its combination of high density and low cost.

YAFFS2 is an acronym for "Yet Another Flash File System, 2nd edition" and provides a high-performance interface between the Linux kernel and NAND flash devices. YAFFS2 was already freely available for Linux. However, it is not part of the standard 2.6.25 Linux kernel, and so Google added it to Android."

From http://nthcode.com/pubs/porting-android-to-a-new-device.html (cheers to Blazeitup123 for posting that link).

[Mr.Wuan]

Can somebody post the link to the latest version to the android on the htc touch pro?

[GlemSom]

Are their any complete SDK kits for Android on HTC Diamond?
(Including kernel, android, cross compiler, binutils and glibc/ulibc?)

Or will I have to start by getting a working crosscompiler platform - and then getting kernel and programs(android) to compile using that toolchain?

[-Mr. X-]

thanx NetRipper


my RAPH800 sprint is getting stuck after:

[26.720054] drivers /rtc/hctosys.c: unable to open rtc dev ice (rtc0)
[27.372075] trout_pwrsink_set: STUB!
[27.379887] trout_pwrsink_set: STUB!
[27.382105] trout_pwrsink_set: STUB!
[27.391098] trout_pwrsink_set: STUB!

maybe set = get txt is small hard tor read but yeah it stuck there any ideas i left it there overnight thinking maybe takes a while fell sleep woke up this morning it was still stuck there.

tested Angstrom for the heck of it also gets stuck same place
but around [21.382075] time
boot from internal memory not sd card.

[dcordes]

Dudes, did you know that TomTom devices also use MoviNAND chips?
And that they have the kernel source diffs downloadable on their website?

Just take a look here: http://www.tomtom.com/page.php?Page=gpl#kernel

Browsing thru the diffs i found a couple of references to MoviNAND functionality like:
+ * at4_sd.c - Cenatrality AT2/3 SDIO Host driver
+ *
+ * Copyright (c) 2007 Centrality Communications, Inc.
+ * Original version for AT4X0A
+ * by Roger Huang <roger.huang@centralitynav.com>
+ * Modified to support MoviNAND and Power Management

I hope that someone can use this for improving the kernel for diamond devices.

what's wrong with the msm_nand ?

[ATT-Half]

boot from internal memory not sd card.

works now strange it did not work when i copied the .exe to my devices folder but it worked when i cut and pasted to internal storage folder.


thanx

[NetRipper]

what's wrong with the msm_nand ?
Just to clearify. In the Diamond there are TWO NAND devices. The first NAND has the SPL, OS, radio and some program space. The second NAND is the MoviNAND which (by the way) is compatible with MMC protocol and contains the "internal storage" of 4gb.

The internal storage (movinand) is the one that should be focused on for Diamond. I doubt that it is the same as msm_nand. In fact, I'm hoping the MoviNAND is just accessible like an SD card via the msm mmc host. We'll know when we get SD card to work on Raphael (or movinand to work on Diamond).

[dcordes]

can somebody try this initrd (http://linuxtogo.org/~lgorris/builds/OE/x11-image-120708.cpio.gz)? it has the angstrom rootfs plus x11 packages. what happens when you boot it? does it display a nice gui when finished?

net, did you try msm_nand on the raphael?

[balsat]

can somebody try this initrd (http://linuxtogo.org/~lgorris/builds/OE/x11-image-120708.cpio.gz)? it has the angstrom rootfs plus x11 packages. what happens when you boot it? does it display a nice gui when finished?

net, did you try msm_nand on the raphael?

It outputs a lot of stuff and then stops with black screen and a white cursor in upper left corner!

[dcordes]

can you move the cursor? did you go through the calibration?

[balsat]

can you move the cursor? did you go through the calibration?


It did not stop for the calibration, but continue for the init... I can't move the cursor, and there is no USB net.

[NetRipper]

can somebody try this initrd (http://linuxtogo.org/~lgorris/builds/OE/x11-image-120708.cpio.gz)? it has the angstrom rootfs plus x11 packages. what happens when you boot it? does it display a nice gui when finished?

net, did you try msm_nand on the raphael?
Well, it doesn't work out of the box. Druidu played with it on his Diamond and you know it almost bricked his device. I stopped playing with it, as it's for the OS part, not for SD. And I don't want to risk bricking my device. We don't need the nand on which OS resides working anyway (low priority at least).

Btw I tried that initrd of yours and on my raph it just gives a black screen when loading X. I don't get a cursor like balsat resports.

[NetRipper]

It did not stop for the calibration, but continue for the init... I can't move the cursor, and there is no USB net.
You can calibrate it anyway. The first tap on the screen inits the calibration, the second tap must be top left, the 3rd tap must be lower right. Do that and see if you can move the cursor. :)

[balsat]

You can calibrate it anyway. The first tap on the screen inits the calibration, the second tap must be top left, the 3rd tap must be lower right. Do that and see if you can move the cursor. :)

I tried that and the cursor (A non blinking white line) still wont move!

[dcordes]

it seems like this is the same problem as on the kaiser where also the first frame drawn by Xfbdev gets stuck.

this proves it's typical on all devices using msm_fb so it's a problem with the driver itself.

[NetRipper]

I tried that and the cursor (A non blinking white line) still wont move!
Oh, that sounds like a hang. I thought you had a graphical cursor. ;)

[dcordes]

retry a few times and see if you can get it to draw something. also you can remove the gpe-dm from init.d and run Xfbdev manually from remote host through usbnet or on a different tty

on the kaiser the display also freezes sometimes before it draws one frame (also on the white cursor line)

[touch of jobo]

Regarding the flickering on screen keyboard..
[...] as i dont see anyone got a workaround with the flickering to FB written soft keyboard [...]

A workaround for the flickering is to put a msmfb_update(registered_fb[0],
MSM_VKEYB_LEFT,
MSM_VKEYB_TOP,
MSM_VKEYB_LEFT + MSM_VKEYB_COLS * MSM_VKEYB_KEY_W,
MSM_VKEYB_TOP + MSM_VKEYB_ROWS * MSM_VKEYB_KEY_H);at the end of msm_vkeyb_draw_keyboard(); (plus a little bit of fiddling as msmfb_update() is static in msm_fb.c and registered_fb[] is extern from fbmem.c) Together with plotting not only on (x,y) but also on (x,y + lcd_height), as I mentioned in this post (http://forum.xda-developers.com/showthread.php?p=2922380#post2922380), you get a non-flickering keyboard with nicely responsive red pressed keys. (You probably only need to draw in 1 of the 2 buffers ... I dont yet know how to tell which one.)

As a side effect, you can then also get rid of the // XXX - VERY ugly hack around line 207 that forces redraw of the entire buffer every time. This breaks the yellow-dot-drawing though. If you value that you have to mark the buffer dirty there too. (Same goes for test pattern drawing.)



The messed up colors thing:
[...] I think nobody find a solution for the color problem though
I've been playing around with the bit packing settings in msm_fb.c This does impact the post handover console drawing and the blitting of the tux image,
http://members.home.nl/j0b0/altered_pixpacking.jpg
but as soon as the android GUI kicks in, everything looks normal. (I have a GSM Raphael, so my colors are OK, I'm _trying_ to get them messed up..)

[...] Bar 03 is the brightest blue, then bar 09 is the brightest green, then bar 13 is the brightest red. [...]
@hillsdale, or anyone with a cdma device: does the tux logo look normal, color wise, and if not, does it if you change to:
red.offset = 10;
red.length = 4;
green.offset = 6;
green.length = 4;
blue.offset = 0;
blue.length = 4;
in msm_fb.c, msmfb_probe(), around line 707 ?



can somebody try this initrd (http://linuxtogo.org/~lgorris/builds/OE/x11-image-120708.cpio.gz)?
I get an error downloading it...

-Jobo

[Untouchab1e]

Ive been following this ever since the first Android images came out, but I am curious as to how far away from a fully functional port we are? I know that regarding Diamond and the Raphael, we still cant even boot it properly, but do you find it likely we will have a fully working version of Android on our HTC phones any time soon?

[n.mowberry]

Ive been following this ever since the first Android images came out, but I am curious as to how far away from a fully functional port we are?

Ages, i actually don't think they'll get it going.

[NetRipper]

Ages, i actually don't think they'll get it going.
Wow, the vote of confidence is overwhelming. ;)

[smile_gerard]

Ages, i actually don't think they'll get it going.

If there is a will.... There is a way :)

[liamhere]

Ages, i actually don't think they'll get it going.


rome wasn't bulit in a day.....give them time

if you really want it now and can't wait.....go buy G1 ;)

[n.mowberry]

Wow, the vote of confidence is overwhelming. ;)

Lawl, sorry just sick of annoying people asking asking asking.

I have confidence in you!!

[petay]

i tried the x11 angstom image and got the same. little white cursor type thing. i saw no calibration but i did notice there were a few errors in the console while it was booting on my gsm diamond. i will try and make notes what they are!!

ps after a few more boots i can not read the messages fast enough to write down the errors, however i have found the calibration is in there. if you wait for the STUB! STUB! STUB! STUB! then click the screen you see the calibration messages within all the other console stuff. i managed to calibrate before getting to the black screen but alas nothing new happens :( i can not move the cursor or anything

[Untouchab1e]

How can I help in porting Android? Is there any donations needed, etc?

[zero2k58]

same question: how can i help you guys without breaking my htc touch diamond?

[slovoflud]

How can I help in porting Android? Is there any donations needed, etc?

same question: how can i help you guys without breaking my htc touch diamond?

Uniting your questions...
You guys can send your diamonds to any of the people who're working on it... Make sure it's packed in Bubble Wrap, so that it won't brake...

[touch of jobo]

can somebody try this initrd (http://linuxtogo.org/~lgorris/builds/OE/x11-image-120708.cpio.gz)?

After the kernel boots, it outputs a lot of stuff to the console. Then it shows (for a very short time) the classic X windows background pattern with the classic X shaped X windows cursor. After that, the screen goes white, in the upper left corner are dark blue cross hairs (plus shaped). Above the center of the screen it says 'touch the crosshairs to calibrate the screen'.

Thats it. Touching the cross hairs does nothing. (Touching the cross hairs some more draws yellow dots on it.) At this point I don't see the X cursor anymore.

It does require a kernel with the hack in msmfb_start_dma() that marks the entire buffer dirty. The kernel still runs, as you can keep drawing dots and the vkeyb still shows pressed keys red. I did go through the calibration process in the touchscreen driver.

(Tested on a GSM Raphael)

-Jobo


EDIT: I shot a movie of it booting (http://members.home.nl/j0b0/boot-linux-x11.mp4). This time it shows the X cursor longer and the background is black, not the classic pattern. Its not quite sharp enough to read clearly but if you know what youre looking for you can probably recognize it. Hope it helps. -Jobo

[freakolowsky]

Uniting your questions...
You guys can send your diamonds to any of the people who're working on it... Make sure it's packed in Bubble Wrap, so that it won't brake...

and other than that ... due to the fact that my Raph is the only working phone i have :D

I'd be glad to pich in with some coding, but due to the lack of time all i'm offering is happy thoughts, money donations, mirror space for testcases on my servers, ...

[brainbubble]

FYI

Android 1.0 SDK 2 is out....
=>http://code.google.com/android/RELEASENOTES.html

[Vince13]

FYI

Android 1.0 SDK 2 is out....
=>http://code.google.com/android/RELEASENOTES.html

Thanks you!

Now we should be able to use the Windows USB Driver in order to debug. It was already possible on UNIX, let's see with Windows now.

[sibbor]

EDIT: I shot a movie of it booting (http://members.home.nl/j0b0/boot-linux-x11.mp4). This time it shows the X cursor longer and the background is black, not the classic pattern. Its not quite sharp enough to read clearly but if you know what youre looking for you can probably recognize it. Hope it helps. -JoboThese kind of videos are great for showing up the progress. I would love to see more videos as we progress further, linked from the Linux/Android for Diamond wiki too.

[touch of jobo]

Regarding the on screen keyboard.. A workaround for the flickering is to put[...]
[...]does the tux logo look normal, color wise, and if not, does it if you change[...]

Anyone who wants to try and see/feel/report what the effect is, you can find a zImage here (http://members.home.nl/j0b0/zImage-jobo-p1027) that..
..has pixel bit packing set to match xxRrrrGg-ggxxBbbb,
(tux may look normal on cdma? or does it already? things are very green / cyan on gsm)
..does not have the hack in msm_fb.c/msmfb_start_dma() that would force updating the entire screen,
..plots on (x,y) and (x,y + ymax),
..calls msmfb_update() in msm_vkeyb_draw()

-Jobo

[jeforce]

Anyone who wants to try and see/feel/report what the effect is, you can find a zImage here (http://members.home.nl/j0b0/zImage-jobo-p1027) that..
..has pixel bit packing set to match xxRrrrGg-ggxxBbbb,
(tux may look normal on cdma? or does it already? things are very green / cyan on gsm)
..does not have the hack in msm_fb.c/msmfb_start_dma() that would force updating the entire screen,
..plots on (x,y) and (x,y + ymax),
..calls msmfb_update() in msm_vkeyb_draw()

-Jobo

I have try your image on a diamond whith the x11-image-120708.cpio.gz , all was green and the dot wasn't in the corner.

[NetRipper]

Anyone who wants to try and see/feel/report what the effect is, you can find a zImage here (http://members.home.nl/j0b0/zImage-jobo-p1027) that..
..has pixel bit packing set to match xxRrrrGg-ggxxBbbb,
(tux may look normal on cdma? or does it already? things are very green / cyan on gsm)
..does not have the hack in msm_fb.c/msmfb_start_dma() that would force updating the entire screen,
..plots on (x,y) and (x,y + ymax),
..calls msmfb_update() in msm_vkeyb_draw()

-Jobo
Could you make a .diff? When using git, just use git-diff > my.diff, no arguments needed.

[touch of jobo]

I have try your image on a diamond whith the x11-image-120708.cpio.gz , all was green and the dot wasn't in the corner.
Is your diamond a GSM one? (Did you already have proper colors in the android background for instance?)
I think that the X server on that initrd image requires a kernel _with_ that msmfb hack because it doesn't mark the frame buffer dirty. This kernel only dma's the dirty area. It is probably better for now to leave that hack in, I guess.

If you have proper colors in android, yet everything goes green on this X server, that would indicate that it does honor the bit packing as set up in msmfb_probe() unlike android. Will test and confirm.

Could you make a .diff? When using git, just use git-diff > my.diff, no arguments needed.
I sure can. Here it is (http://members.home.nl/j0b0/linux-htc-msm-2.6.25-jobo20081208.diff.txt). Its a bit big because of tinkering around (did some cleaning up on that now..) and its not against the latest git.

-Jobo


EDIT:
The X server does indeed draw green/blue when I set the bit packing differently in msmfb_probe(). However, it only draws the X cursor for a second or so and then drops back to the console. I get a login prompt. I assume the X server crashes. When I put the bit packing back to 'correctly' match rgb565, the X server stays up and draws the cross hairs and the instruction to touch it, as in the movie (http://members.home.nl/j0b0/boot-linux-x11.mp4).
red offset = 11, length = 5;
green offset = 5, length = 6;
blue offset = 0, length = 5;
Again, the X server needs the XXX - very ugly hack (msm_fb.c~207) in place.

-Jobo

[Untouchab1e]

How will the Android SDK 2 affect the porting? Will it help in any way, or will it just allow much more to be done from Windows rather than Linux?

[maejrep]

@hillsdale, or anyone with a cdma device: does the tux logo look normal, color wise, and if not, does it if you change to:
red.offset = 10;
red.length = 4;
green.offset = 6;
green.length = 4;
blue.offset = 0;
blue.length = 4;
in msm_fb.c, msmfb_probe(), around line 707 ?

No, my colors do not look normal with the git kernel, but your changes DO make it look normal! So what does that mean? We do have 3 different displays?

Also worth noting that current git doesn't boot on my phone anymore. I think I'm going to revert the df6bc3917c80b8e4e6d718d58e4ca227259a8408 patch for usb ether and see if that's the cause. The symptom is the same as others are reporting on the previous page, just hangs after the 4 STUB lines, no chance to calibrate.

[jmacdonald801]

Sure would be nice if HTC, or Quallcomm was a little more forthcoming about there hardware. Talk about throwing a dog a bone... Open Handset Alliance? What alliance. lol.

The point to be made here, is that it's pretty sad that open source devs are going it alone when there is all this talk about the open source alliance, but regular joe's are having to do all the work.

Guess it just pisses me off, after having bought a few HTC phones. Everyone was pissed about the Kaiser not having proper video drivers, and again, where are these companies other than spouting off propaganda?

Sorry I have nothing positive to add here. Really wish there was something I could contribute.

Sincere thanks, for doing what the billion dollar companies can't spare a few dollars to do themselves.

-James

[sibbor]

No, my colors do not look normal with the git kernel, but your changes DO make it look normal! So what does that mean? We do have 3 different displays?

Also worth noting that current git doesn't boot on my phone anymore. I think I'm going to revert the df6bc3917c80b8e4e6d718d58e4ca227259a8408 patch for usb ether and see if that's the cause. The symptom is the same as others are reporting on the previous page, just hangs after the 4 STUB lines, no chance to calibrate.Great to hear... sounds like we got some neat progress here :D

[zonyl]

Unfortunately I dont have much time to help, however, I am excited about the efforts of getting Linux / X running and have money if that helps. Is there a specific link to donate to the group working on this (post on the Wiki page please)? If one cannot be organized, could those folks who are all contributing here put a donate link in their sigs? Thanks for all of your efforts!!

[dcordes]

After the kernel boots, it outputs a lot of stuff to the console. Then it shows (for a very short time) the classic X windows background pattern with the classic X shaped X windows cursor. After that, the screen goes white, in the upper left corner are dark blue cross hairs (plus shaped). Above the center of the screen it says 'touch the crosshairs to calibrate the screen'.

Thats it. Touching the cross hairs does nothing. (Touching the cross hairs some more draws yellow dots on it.) At this point I don't see the X cursor anymore.

It does require a kernel with the hack in msmfb_start_dma() that marks the entire buffer dirty. The kernel still runs, as you can keep drawing dots and the vkeyb still shows pressed keys red. I did go through the calibration process in the touchscreen driver.

(Tested on a GSM Raphael)

-Jobo


EDIT: I shot a movie of it booting (http://members.home.nl/j0b0/boot-linux-x11.mp4). This time it shows the X cursor longer and the background is black, not the classic pattern. Its not quite sharp enough to read clearly but if you know what youre looking for you can probably recognize it. Hope it helps. -Jobo


That's interesting. I'm glad some more people tested it. I wonder why do some experience the freeze when X attempts to start and you get to the calibration screen. But I think it's not dependend on the mddi client because on the kaiser it sometimes draws one frame and then freezes, sometimes even freezes before it draws anything. On the other hand you are the only person who seem to get anything at all drawn to the fb by X. Did you ever get the black screen only without any cursor?

If you want to sort out it's not just the touchscreen driver not working, you can login and run Xfbdev manually, then run some application on it with 'DISPLAY=:1 pidgin' or so. Just remove the gpe-dm from init. Would also be interesting to see if druidu's ts driver works with tslib so you might also run tslib_test tslib_calibrate.

[skiwithpete]

Looks like the HTC touch is android-ed.

http://www.engadget.com/2008/12/10/video-android-rocking-on-the-htc-touch/

[kragil]

Looks like the HTC touch is android-ed.

http://www.engadget.com/2008/12/10/video-android-rocking-on-the-htc-touch/

Wow, super cool!

I want Android on my Diamond and I am willing to put my money where my mouth is.
I will pledge 10$ (for a bounty) to get Android running that well on the Diamond. Anyone willing to join in? If enough people join in we might get a decent sum together.

[Stebo85]

please guys i cant wait to have android because winmo just sux. been using it for 3 moths and its driving me "BLEEPing" crazy! im ready to pay an contribution for this.

keep me updated

[courtlandre]

Wow, super cool!

I want Android on my Diamond and I am willing to put my money where my mouth is.
I will pledge 10$ (for a bounty) to get Android running that well on the Diamond. Anyone willing to join in? If enough people join in we might get a decent sum together.

It has been running pretty stable like this for over a month. WinMo is still better because it actually supports apps. Android on the vogue still has many problems and in my opinion is not in a usable state.

[Untouchab1e]

Would also be willing to pay whatever it takes to get Android running properly on the Touch Pro :D Please let me know if there is anything I can do!

Keep up the great work!

[Mikey]

please guys i cant wait to have android because winmo just sux. been using it for 3 moths and its driving me "BLEEPing" crazy! im ready to pay an contribution for this.

keep me updated

You are not on a playground.
Please watch your language.

[garuda50]

What a long treat yessurr!!! You guys are desperately convinced that u could make a HTC diamond run on Android?? Whuhahahahah don't make me laugh:mad:

Instead of wasting your time on your precious girl.. Guys or should i say d....

Use your heads! I know the truth comes very hard but please go work with google to make an real OS but U bunch of amateurs.

Please close this treat MODERATORS.. This is just cowsh.t...

:mad: Im not pleased to see a new reply that goes in the middle of nowhere.

Now people, if u have a mind and have the balls. We should stop this stupid threat. Think forward not with a bow. I know im right!!!:mad:

For shizzle!! Yessurr!!

[Mikey]

The next posting you make better be to someone that already posted or you will be placed in the time-out corner. :D
Got my drift??

[garuda50]

I just dont want people to donate them wannabe programmers and developers. Just because one silly man asked if its possible to make an google OS run on Diamond.

It's like if u ask.. Does Santa Claus really excist? No it don't excist but we make them excist.

Im for real people. PLEASE STOP this shizzle. Im really mad about the progressing here. I dont apreciate the work done here cause it leads to nowhere. PLEASE stop ok! Youre making other people desperate :rolleyes:

[garuda50]

You think you tough huh?!?! Come and meet me cowboy!!!:mad:

[Mikey]

I just dont want people to donate them wannabe programmers and developers. Just because one silly man asked if its possible to make an google OS run on Diamond.

It's like if u ask.. Does Santa Claus really excist? No it don't excist but we make them excist.

Im for real people. PLEASE STOP this shizzle. Im really mad about the progressing here. I dont apreciate the work done here cause it leads to nowhere. PLEASE stop ok! Youre making other people desperate :rolleyes:

These forums are open to open topics. Whether something is possible or eventually possible.

Are you so definite that Android wont be able to run on the Pro?

[badbrad4080]

Yes mikey, im for shizzle about it. Look hardware its ok about it. But the drivers are not support to go with the internal shizzle of the DIAMOND. Please dont let it go to yor head.. Or you believe cows can fly:rolleyes:

LMao i can tell you know nothing about software mod that timeout would be good now and through ip give his other accounts timeout.

[Mikey]

in about 1 minute sign in with another account :D

[garuda50]

GIVE ME time out then, i have 4 other accounts brother. Yeah you know nuthin bout it bout naah mean... Go buy yourself a HTC g1 if u ask santa maybe he'll give u one!!!:D

[abovelegit]

GIVE ME time out then, i have 4 other accounts brother. Yeah you know nuthin bout it bout naah mean... Go buy yourself a HTC g1 if u ask santa maybe he'll give u one!!!:D

Im pretty sure he can ban your IP adress just as easily.

[Mikey]

Im pretty sure he can ban your IP adress just as easily.

I have all his names :)
No need to ban them all. Maybe he has a different personality with each name he has. lol

[Kasaki]

Great going so far guys, iI quite developing stuff a while back but this looks like great progress. Now the issue with the cdma phones and the android os is? excuse if if the question has been asked. Also the android os never came with a virtual keyboard to begin with correct?

[muscleman7911]

I have all his names :)
No need to ban them all. Maybe he has a different personality with each name he has. lol

I come from no wer into this topic but its amusing,
I see it that this personality is just mad that Android is not on his diamond or raphael... yet..
LOL:D:D:D

[maejrep]

No, my colors do not look normal with the git kernel, but your changes DO make it look normal! So what does that mean? We do have 3 different displays?

Also worth noting that current git doesn't boot on my phone anymore. I think I'm going to revert the df6bc3917c80b8e4e6d718d58e4ca227259a8408 patch for usb ether and see if that's the cause. The symptom is the same as others are reporting on the previous page, just hangs after the 4 STUB lines, no chance to calibrate.

So going over that diff (http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commitdiff;h=df6bc3917c80b8e4e6d718d5 8e4ca227259a8408), I found an issue that may have already been brought up or addressed, but from what I can tell, latest git is still wrong..

.ifc_class = 0x02,
#ifdef MACH_HTCRAPHAEL
.ifc_subclass = 0x00
#else
.ifc_subclass = 0x0a,
#endif
.ifc_protocol = 0x00,

If MACH_HTCRAPHAEL is set, you end up with:
.ifc_class = 0x02,
.ifc_subclass = 0x00
.ifc_protocol = 0x00,

Which is saying:

.ifc_class = 0x02,
.ifc_subclass = 0x00.ifc_protocol = 0x00,

That should probably be fixed in git. I was hoping that might be the reason for the hangs that we're seeing on the raph800, but even after fixing that, I still hang at the same spot. I'll try fresh and see if maybe I screwed something up at some point.

[Arktronic]

.ifc_class = 0x02,
.ifc_subclass = 0x00.ifc_protocol = 0x00,That should probably be fixed in git. I was hoping that might be the reason for the hangs that we're seeing on the raph800, but even after fixing that, I still hang at the same spot. I'll try fresh and see if maybe I screwed something up at some point.

Shouldn't that kind of mistake have caused a compiler error? You wouldn't even be able to build an image then. Perhaps MACH_HTCRAPHAEL isn't being defined properly (if at all)?

[maejrep]

Shouldn't that kind of mistake have caused a compiler error? You wouldn't even be able to build an image then. Perhaps MACH_HTCRAPHAEL isn't being defined properly (if at all)?

Indeed, you should get "error: request for member 'ifc_protocol' in something not a structure or union". How it did manage to compile is beyond me, unless like you said, MACH_HTCRAPHAEL wasn't being defined, which checking now, is clearly the case:

CC drivers/usb/function/ether.o
drivers/usb/function/ether.c:217:2: warning: #warning Not Set

#ifdef MACH_HTCRAPHAEL
#warning Is Set
.ifc_subclass = 0x00,
#else
#warning Not Set
.ifc_subclass = 0x0a,
#endif

Looks like that is intended to be #ifdef CONFIG_MACH_HTCRAPHAEL

[NetRipper]

Indeed, you should get "error: request for member 'ifc_protocol' in something not a structure or union". How it did manage to compile is beyond me, unless like you said, MACH_HTCRAPHAEL wasn't being defined, which checking now, is clearly the case:

CC drivers/usb/function/ether.o
drivers/usb/function/ether.c:217:2: warning: #warning Not Set

#ifdef MACH_HTCRAPHAEL
#warning Is Set
.ifc_subclass = 0x00,
#else
#warning Not Set
.ifc_subclass = 0x0a,
#endif

Looks like that is intended to be #ifdef CONFIG_MACH_HTCRAPHAEL
Doesn't this mean the #ifdef is actually not needed at all? I can ssh to my Raphael with the current code, thus it works with the 0x0a subclass. Let me know.

[konsta819]

I'm sure the gurus are aware of this, but the Vogue project met with considerable success today - they are now running Android 1.0 with everything working. Is there anything that can be learned from their devs?

http://www.engadget.com/2008/12/10/video-android-rocking-on-the-htc-touch/

http://forum.xda-developers.com/showthread.php?t=382265

[maejrep]

Doesn't this mean the #ifdef is actually not needed at all? I can ssh to my Raphael with the current code, thus it works with the 0x0a subclass. Let me know.

That's probably true, but lavender.t said he had problems with the x0a subclass. Could be only a certain phone (GSM/CDMA, Diamond/Raphael, ..) that has that problem.

And as we know, booting on one device doesn't mean it'll boot on all :) We might need to find out from lavender.t what exactly happens with 0x0a on his phone, etc.

[touch of jobo]

No, my colors do not look normal with the git kernel, but your changes DO make it look normal! So what does that mean? We do have 3 different displays?
It means that your LCD wants pixel data in the frame buffer in xxRrrrGg-ggxxBbbb format, like hillsdale suggested in post #896 (http://forum.xda-developers.com/showthread.php?p=2940282).

[...] But I think it's not dependent on the mddi client because on the kaiser it sometimes draws one frame and then freezes, sometimes even freezes before it draws anything.

I will try a couple more times to confirm, but it felt very much like changing the bit packing made the X server quit or crash. Will get back on that.
During the 1 second that the X cursor is there, its color is affected.

On the other hand you are the only person who seem to get anything at all drawn to the fb by X. Did you ever get the black screen only without any cursor?

When I removed the hack from msm_fb that dma's the entire screen every time, I don't see X draw anything. Other than that it has always started up and drawn.

-Jobo

[Willis111]

When I removed the hack from msm_fb that dma's the entire screen every time, I don't see X draw anything. Other than that it has always started up and drawn.
-Jobo

I was just wondering which initrd you are using to boot. Could you post it/host it/link it or whatever?

EDIT: With the x11 initrd and initrd-android-raphael.cpio.gz 05-Dec-2008 it's stopping for me at the framebuffer handover. Using a cdma telus touch diamond.

[Hamido123]

sorry for this post my keyboard getting errors

[Hamido123]

hallo to everybody,
i am one of many people who want to get android 1.0 running on the diamond and i want to develop with you but i don`t know why android doesn´t work and how progressed you are so far so please can someone post a short "diary of android for diamond"


i have read in archives that we were not able to boot android because we do not got a SD card is that true and if it is so why we do not rename in the registry the attributs of internal storage as we want to?


sorry for my bad english i am a kurd

[touch of jobo]

I was just wondering which initrd you are using to boot. Could you post it/host it/link it or whatever?

It is this one (http://linuxtogo.org/~lgorris/builds/OE/x11-image-120708.cpio.gz), the one dcordes requested to test and report back on in post #1017 (http://forum.xda-developers.com/showthread.php?p=3010181#post3010181).


EDIT: With the x11 initrd and initrd-android-raphael.cpio.gz 05-Dec-2008 it's stopping for me at the framebuffer handover. Using a cdma telus touch diamond.

That would have more to do with your kernel/zImage than with your initrd. At that point it hasnt started decompressing the initrd yet.

-Jobo

[Willis111]

please can someone post a short "diary of android for diamond"



http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux

[touch of jobo]

I will try a couple more times to confirm, but it felt very much like changing the bit packing made the X server quit or crash. Will get back on that.

I have tested another bunch now. Using dcordes x11 initrd and 2 almost identical kernels. Only difference is the 444 vs. 565 bit packing definition in msmfb_probe(). The 565 (link to kernel image) (http://members.home.nl/j0b0/zImage-565-jobo) one _always_ makes it all the way through (link to that same movie) (http://members.home.nl/j0b0/boot-linux-x11.mp4) to the 'touch the crosshairs to calibrate'-message. The 444 (link to kernel image) (http://members.home.nl/j0b0/zImage-444-jobo) one _always_ shows the X cursor for a second and then drops back (link to new movie that shows this) (http://members.home.nl/j0b0/boot-linux-x11-444.mp4) to the login prompt.
When I then go through the msm_ts calibration process, login and start the X server, it stays up but doesn't show the crosshairs or the message. Just an X cursor on an X background:
http://members.home.nl/j0b0/blueX.jpg
I can not move the cursor. ..or do anything at all for that matter. ..except of course drawing yellow dots.

If you want to sort out it's not just the touchscreen driver not working, you can login and run Xfbdev manually, then run some application on it with 'DISPLAY=:1 pidgin' or so. Just remove the gpe-dm from init. Would also be interesting to see if druidu's ts driver works with tslib so you might also run tslib_test tslib_calibrate.
init links to /sbin/init.sysvinit which is an ELF binary. But for logging in and launching stuff, right now, booting the '444 kernel' has the same effect...
Those tslib tools are not in the fs image (nor is gpe-dm). Can I find them somewhere or cross compile from source and add to the initrd or can you put together one that includes those tools? I'd be happy to run tests and report back.

-Jobo

[Willis111]

I have tested another bunch now. 565 (link to kernel image) (http://members.home.nl/j0b0/zImage-565-jobo)
444 (link to kernel image) (http://members.home.nl/j0b0/zImage-444-jobo)



Both of those flop out @ "console handover: boot[htc_fb0] -> real [tty0]"
2 messages before that I see "console [htc_fb0] enabled" which I would assume is a good thing.

Of course, I'm using a cdma touch diamond and assuming that you're running the gsm version of the same phone.

[nuke1]

Both of those flop out @ "console handover: boot[htc_fb0] -> real [tty0]"
2 messages before that I see "console [htc_fb0] enabled" which I would assume is a good thing.

Of course, I'm using a cdma touch diamond and assuming that you're running the gsm version of the same phone.
CDMA is probably different enough to warrant lauching a new thread dedicated to that solely. :p

[NetRipper]

Both of those flop out @ "console handover: boot[htc_fb0] -> real [tty0]"
2 messages before that I see "console [htc_fb0] enabled" which I would assume is a good thing.

Of course, I'm using a cdma touch diamond and assuming that you're running the gsm version of the same phone.
The .diff that touch of jobo posted lately didn't include the CDMA mddi panel code. It could be that he didn't apply the latest git to his code yet - thus being the reason it doesn't boot on CDMA.

@touch of jobo, I'm referring to this (http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commit;h=2f981bbd7f1023c7c433c4e666e9 882f22e9180c) commit. If you apply that, the panel is recognized on CDMA devices.

[NetRipper]

Are you guys almost done with this cause my htc pro is in need of the andriod shit on it :)
No, see wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux).

[touch of jobo]

The .diff that touch of jobo posted lately didn't include the CDMA mddi panel code. It could be that he didn't apply the latest git to his code yet - thus being the reason it doesn't boot on CDMA.
@touch of jobo, I'm referring to this (http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commit;h=2f981bbd7f1023c7c433c4e666e9 882f22e9180c) commit. If you apply that, the panel is recognized on CDMA devices.
That is correct. Thanks. I shall apply it and update those 2 kernel images. Probably not until later tonight though..
-Jobo

[sensorama]

Have you guys seen this
http://www.youtube.com/watch?v=6iKje_0lUYg android in HTC Touch

Yes.

in post #998

[NetRipper]

Hi!
I have a problem for start Android on HTC Diamond...Read the FAQ and Progress sections @ wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux). Android is not functional yet.
noob? or are trolling with us? read the F***ing thread!Let's just redirect people to wiki. No point calling names.

[col Cyber]

Read the FAQ and Progress sections @ wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux). Android is not functional yet.
Let's just redirect people to wiki. No point calling names.


Thanks!
:)

you are soo kind

[touch of jobo]

That is correct. Thanks. I shall apply it and update those 2 kernel images. Probably not until later tonight though..
I updated them (because I said I would, not because it is particularly useful, as I have come a bit further in the mean time). Same links as before:
zImage-444-jobo (http://members.home.nl/j0b0/zImage-444-jobo)
zImage-565-jobo (http://members.home.nl/j0b0/zImage-565-jobo)

[...] This does impact the post handover console drawing and the blitting of the tux image, but as soon as the android GUI kicks in, everything looks normal. (I have a GSM Raphael, so my colors are OK, I'm _trying_ to get them messed up..) [...]
Right .. so I was trying...
http://members.home.nl/j0b0/messed-up-colors-on-r100.jpg
..something like that. :)

I got there by playing with:
/* 666 18BPP */
dma2_cfg |= DMA_DSTC0G_6BITS | DMA_DSTC1B_6BITS | DMA_DSTC2R_6BITS;

..in mdp.c~235. So this may be what we need to adjust according to the raph800 lcd.

In order to have different flags for different panels, I did some small (but probably considered ugly and dirty) modifications to mddi_client_toshiba.c, msm_fb.h and mdp.c
First I applied the 'cdma patch' to register the raph800 panel, as netripper pointed out (http://forum.xda-developers.com/showthread.php?p=3031724).

- Changed the prototype of mdp_dma_to_mddi() to have an argument that holds extra flags. These get ORred to dma_cfg2.
- Added a member to the mddi_panel_info struct to hold those flags per panel
- Added an enable function for the raph100 panel, so that they now both have one. In here, each panel sets the extra mdp dma flags it needs. msmfb_start_dma() passes these to mdp_dma_to_mddi()
- Flags that should be controlled on a per panel basis removed from mdp_dma_to_mddi() #ifdef CONFIG_MACH_HTCRAPHAEL
In mddi_client_toshiba.c:
static void toshiba_enable_panel_raph800(struct mddi_panel_info *panel)
{
panel->width = 480;
panel->height = 640;
/* Jobo: See which combination works, it does seem to matter */
panel->mdpdmacfgdstflgs = DMA_DSTC0G_5BITS | DMA_DSTC1B_5BITS | DMA_DSTC2R_5BITS;
panel->mdpdmacfgdstflgs |= DMA_PACK_TIGHT; // or DMA_PACK_LOOSE
panel->mdpdmacfgdstflgs |= DMA_DITHER_EN; // or not
}

static void toshiba_enable_panel_raph100(struct mddi_panel_info *panel)
{
/* Jobo: Set values to what use to be fixed, width/height already OK or not needed */
panel->mdpdmacfgdstflgs = DMA_DSTC0G_6BITS | DMA_DSTC1B_6BITS | DMA_DSTC2R_6BITS;
panel->mdpdmacfgdstflgs |= DMA_PACK_TIGHT | DMA_DITHER_EN;
}

In mdp:c~235:

#ifdef CONFIG_MACH_HTCRAPHAEL
/* Jobo: add flags from mddi_panel_info */
dma2_cfg |= dmacfgdst;
#else
dma2_cfg |= DMA_PACK_TIGHT | DMA_DITHER_EN;
/* 666 18BPP */
dma2_cfg |= DMA_DSTC0G_6BITS | DMA_DSTC1B_6BITS | DMA_DSTC2R_6BITS;
#endif

This way different panels can use different dma flags in mdp while keeping mdp itself panel agnostic.

You can find a kernel image here (http://members.home.nl/j0b0/zImage-mdptest-jobo) that has these changes. It has the pixel bit packing in msm_fb.c set to the original rgb565 values.

All 3 kernels linked to from this post have the msmfb_start_dma() hack in place, so this should not be the cause of stuff not getting drawn. All 3 have the patch applied that I forgot earlier, so raph800 panels should get recognized.

The raph800 sets different mdp dma dst flags. I would be very interested to know what it looks like. If for some reason my kernels dont run on CDMA phones, I would still be interested what the effect is of different combinations of DMA_DSTCXX_XBITS in dma2_cfg in mdp_dma_to_mddi() mdp.c~235 and maybe also loose or tight packing, lsb or msb aligned, dither enabled or not.

I should probably update to the latest git in order to produce clean diffs. How easy and/or safe is it to merge more recent commits into an older set with modifications?

-Jobo

EDIT: I updated the mdptest image on 12/14 00:45. please boot it.

[lavender.t]

And as we know, booting on one device doesn't mean it'll boot on all :) We might need to find out from lavender.t what exactly happens with 0x0a on his phone, etc.

sorry just saw this post for being busy for a while. yes 0x0a works too but it just made the host system thinking the phone was talking MDLM and gave annoying warnings (because the ether implemented was indeed CDC Subset, which doesnt have a proper subclass yet). and i was worried that somewhere down the road it might have some trouble. that's the reason i changed it to 0.

i kinda doubt this was the reason that your phone didnt boot. but if you change it back to 0x0a and your problem is gone then feel free to revert this file.

[schweinshax]

Hi!
I have a problem for start Android on HTC Diamond...
I download android-05-12-08.zip
i uncompress it and copy it on the internal memory (4gb) in folder android.
I start the haret.exe
i try to connect... nothing...
ok...
i try to start..
haret: booting linux ...
ok... i wait...
wait...
wait...
uhm... freeeeaze!
:(
Why?

I try with
set MTYPE 1439
and
set MTYPE 1667

I dont found a guide for install android.

Can you help me?

Thanks! :D

you need more free space on your internal drive

[Black_Angel]

Hi everyone,

Sorry I don't speak english very well...

So, yesterday, I ran Android on my HTC TD and it succesfully boot. But, on the "home screen" I don't know what to do :eek:
It seems the touchscreen don't work very well, if I take the stylus, I arrive on a contact page and stars appears.
That's all I can do.

Does anyone know what can I do with this version of Android on my phone?

Thnks :)

[cobain84]

I'm looking forward to getting a RAPH800 and helping out with the dev. Either way if the current devs would like I have MASSIVE amounts of server space I do not use that I would be happy to donate to this project for website/blog/CVS. I'd like to see the Raph be the secord (or third??) Android handset.

[lavender.t]

wonder if anyone has a pointer to how the A0/A4 clock values are calculated. i saw the table in clock-7x00.c but that is pressumably based on the 400MHz clock, which is different from the diamonds 528MHz.

also, why is there a 'freq' and an 'actual_freq' which is most half the value of the former ? from the code apparently if function pc_clk_set_rate() is called to set the clock to 2000000, then calling pc_clk_get_rate() would return 144000, etc. i'm quite confused.

i pdumped my diamond and got values A0 = 0001fff4 and A4 = fff50044, matching none of the entries in the table. dont know where these values should fit in (with the freq and actual_freq). and how to get the rest of the entries ... :confused:

//sigh ... i guess probably the biggest help would be getting the datasheet of this MSM chip. but i know what a greedy company qualcomm is. :mad:

- lt

[skiwithpete]

//sigh ... i guess probably the biggest help would be getting the datasheet of this MSM chip. but i know what a greedy company qualcomm is. :mad:

- lt

This is from a post on page29:

For ARM datasheets you can refer to this one, which seems to be most alike with the ARM core in the MSM7201A:
http://infocenter.arm.com/help/topic...01g/index.html

Then further down:
Actually, the exact datasheet for the cpu core is this one,
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ddi0211j/index.html
but this is general -- it won't give us the specifics of MSM7201A like how to drive the MDDI inside the chip or list of registers and memory locations.


I remember that someone had tried to start a map, but I can't find the right post at the minute.

P

[dejant_9]

dejant_9, it would be wonderful if you'll make your forum in French too :)

Merci :D

heuuuu possible but i think that more French speak English than English speak French...first i want to know if everyone here is interested by a forum. I don't want to spend my time for a unused forum

[WatskeBart]

Linux will not boot if you have the usb cable connected. It will say that ARMv9 has crashed and it stays in that screen. Also no keyboard available.
I don't know if this info helps but i just wanted to let you dev's know. :)

If there is anything i can do to help with the development let me know, but do know that i have very little linux expierience.

[jlewis11]

So since android is almost completely running on the Touch, all except for the Voice radio part, do you think we are remotely that close for For Diamond and Touch Pro. You would think that since G1 and Touch Pro are very similar it would be easier...but obviously thats not the case

[touch of jobo]

Here is a post that has something to do with running linux on an HTC Diamond or Raphael phone ;)

[...] If you want to sort out it's not just the touchscreen driver not working, you can login and run Xfbdev manually, then run some application on it with 'DISPLAY=:1 pidgin' or so. Just remove the gpe-dm from init. Would also be interesting to see if druidu's ts driver works with tslib so you might also run tslib_test tslib_calibrate.

The X server runs and draws fine. I can't add another host to the access list with xhost. It restarts the server with a clean list. However, I can ssh in, kill everything, and then relaunch X without access control. Then I can have X client apps run on my linux vm and draw on the raphael.
Input from the vkeyb goes to the focussed X window. (I can type commands in an xterm that runs on my linux vm.)

When I start x-window-manager, the X shaped cursor changes to an arrow pointing NNW. This wm draws some window decoration and also changes the geometry of any client app to full screen, ignoring any -geometry supplied.
When I kill the wm, leaving X running, I can position X clients -geometry.

I can not move the pointer, but I do have a /dev/input/mouse0 that seems to work. When I cat /dev/input/mouse0 and then drag across the screen, I see stuff spewed. It appears to work like a trackpad tho, outputting speed and direction, rather than position. If I boot into netrippers busybox initrd, I have a /dev/mouse0 in stead of a /dev/input/mouse0. The init in that one doesn't seem to explicitly create this device. I can not read/edit the init in the x11 angstrom initrd (or I don't know where to look).

Some photographs:

With WM:http://members.home.nl/j0b0/x11+wm.jpg Without WM:http://members.home.nl/j0b0/x11-wm.jpg

Decoration:
http://members.home.nl/j0b0/x11+wmdeco.jpg

-Jobo

PS: Can someone boot this kernel (http://members.home.nl/j0b0/zImage-mdptest-jobo) on a cdma phone? Thanks.

[pixelwix]

this what i got when i boot it up on my sprint touch pro
colors still messed up and it gets stuck

[Willis111]

PS: Can someone boot this kernel (http://members.home.nl/j0b0/zImage-mdptest-jobo) on a cdma phone? Thanks.

Booting on my telus diamond touch/x11-initrd it boots to the screen calibration at which point clicking the xhairs has no effect and the device seems stalled. The screen also has a bit of a greenish/blueish tint to it. Images attached, first 3 w/ x11-initrd and last 1 w/netrippers initrd(img00049). Tried it w/ netrippers initrd-android-raphael and I'm able to calibrate the screen, get x loaded and play with apps in x though the colors are messed ofc.

@Pixelwix: are you booting this off your internal storage or off your program memory? Try booting off your program memory as it stalls out @ that point every time if you're booting off the internal storage for some reason.

[lavender.t]

i had to use my best guess to crunch the A0/A4 numbers and then out of sudden saw a big step. if the debug log didnt lie to me, the diamond seems to have the moviNAND on SDC2. i got the device /dev/mmcblk0 created and part of the MMC commands ran successfully. although if i "dd if=/dev/mmcblk0" it still times out.

got the following logs:

[ 5.678819] mmc0: new MMC card at address 0001
[ 5.680091] device: 'mmc0:0001': device_add
[ 5.700396] PM: Adding info for mmc:mmc0:0001
[ 5.700427] bus: 'mmc': add device mmc0:0001
[ 5.700641] bus: 'mmc': driver_probe_device: matched device mmc0:0001 with driver mmcblk
[ 5.700671] bus: 'mmc': really_probe: probing driver mmcblk with device mmc0:0001
[ 5.710457] mmcblk0: mmc0:0001 XEE0AM 3932160KiB
[ 5.710518] device: 'mmcblk0': device_add
[ 5.710854] PM: Adding info for No Bus:mmcblk0
[ 5.731190] mmcblk0:<7>mmc0: starting CMD18 arg 00000000 flags 000000b5
[ 5.740152] mmc0: blksz 512 blocks 8 flags 00000a00 tsac 15 ms nsac 1000
[ 5.760122] mmc0: CMD12 arg 00000000 flags 0000049d
[ 5.760183] +msmsdcc_request
[ 5.760213] MSMSDCC: poll=0
[ 5.760244] mmc0: msmsdcc_start_command: op 12 arg 00000000 flags 000000b5
[ 5.760305] mmc0: msmsdcc_irq: irq0 00082040
[ 5.760335] mmc0: msmsdcc_irq: irq0 00082000
[ 5.760824] mmc0: msmsdcc_start_command: op 0c arg 00000000 flags 0000049d
[ 5.760885] mmc0: msmsdcc_irq: irq0 00400040
[ 5.760915] mmc0: req done (CMD18): 0: 00000900 0f5903bf f6dbffe7 964040b8
[ 5.760946] mmc0: 4096 bytes transferred: 0
[ 5.760976] mmc0: (CMD12): 0: 00000b00 0f5903bf f6dbffe7 964040b8
[ 5.761037] mmc0: msmsdcc_irq: irq0 00400000
[ 5.785157]
[ 5.785767] driver: 'mmc0:0001': driver_bound: bound to device 'mmcblk'
[ 5.790091] bus: 'mmc': really_probe: bound device mmc0:0001 to driver mmcblk

the line "[ 5.710457] mmcblk0: mmc0:0001 XEE0AM 3932160KiB" suggested a 4GB MMC was online. there is only one 4GB block, i think. :)

the dd failure logs are:

[ 43.243570] mmc0: starting CMD18 arg 00000000 flags 000000b5
[ 43.250244] mmc0: blksz 512 blocks 32 flags 00000a00 tsac 15 ms nsac 1000
[ 43.262441] mmc0: CMD12 arg 00000000 flags 0000049d
[ 43.270946] +msmsdcc_request
[ 43.281465] MSMSDCC: poll=0
[ 43.289613] mmc0: msmsdcc_start_command: op 12 arg 00000000 flags 000000b5
[ 43.290091] mmc0: msmsdcc_irq: irq0 00400048
[ 43.290091] mmc0: Data timeout
[ 43.290091] mmc0: msmsdcc_irq: irq0 00400000
[ 43.290244] mmc0: DMA channel flushed (0x80000004)
[ 43.298087] mmc0: DATACNT = 16384, FIFOCNT = 0, STATUS = 0x00400000
[ 43.300091] Flush data: 0000c003 1562e000 00000000 00400040 00400040 00000003
[ 43.300091] mmc0: msmsdcc_start_command: op 0c arg 00000000 flags 0000049d
[ 43.300152] mmc0: msmsdcc_irq: irq0 00400040
[ 43.308118] mmc0: req done (CMD18): 0: 00000900 0f5903bf f6dbffe7 964040b8
[ 43.310091] mmc0: 0 bytes transferred: -110
[ 43.310091] mmc0: (CMD12): 0: 00000b00 0f5903bf f6dbffe7 964040b8
[ 43.310091] mmc0: msmsdcc_irq: irq0 00400000
[ 43.321495] mmcblk0: error -110 transferring data
[ 43.330610] end_request: I/O error, dev mmcblk0, sector 0
[ 43.338789] Buffer I/O error on device mmcblk0, logical block 0
[ 43.340091] end_request: I/O error, dev mmcblk0, sector 8
[ 43.340091] Buffer I/O error on device mmcblk0, logical block 1
[ 43.340091] end_request: I/O error, dev mmcblk0, sector 16
[ 43.340091] Buffer I/O error on device mmcblk0, logical block 2
[ 43.340091] end_request: I/O error, dev mmcblk0, sector 24
[ 43.340091] Buffer I/O error on device mmcblk0, logical block 3
[ 43.350336] mmc0: starting CMD18 arg 00000000 flags 000000b5
[ 43.360213] mmc0: blksz 512 blocks 8 flags 00000a00 tsac 15 ms nsac 1000
[ 43.373967] mmc0: CMD12 arg 00000000 flags 0000049d
[ 43.380915] +msmsdcc_request
[ 43.391007] MSMSDCC: poll=0
[ 43.399796] mmc0: msmsdcc_start_command: op 12 arg 00000000 flags 000000b5
[ 43.400091] mmc0: msmsdcc_irq: irq0 00400048
[ 43.400091] mmc0: Data timeout
[ 43.400091] mmc0: msmsdcc_irq: irq0 00400000
[ 43.400305] mmc0: DMA channel flushed (0x80000004)
[ 43.409033] mmc0: DATACNT = 4096, FIFOCNT = 0, STATUS = 0x00400000
[ 43.410091] Flush data: 0000c003 1562e000 00000000 00400040 00400040 00000003
[ 43.410091] mmc0: msmsdcc_start_command: op 0c arg 00000000 flags 0000049d
[ 43.410152] mmc0: msmsdcc_irq: irq0 00400040
[ 43.419277] mmc0: req done (CMD18): 0: 00000900 0f5903bf f6dbffe7 964040b8
[ 43.420091] mmc0: 0 bytes transferred: -110
[ 43.420091] mmc0: (CMD12): 0: 00000b00 0f5903bf f6dbffe7 964040b8
[ 43.420091] mmc0: msmsdcc_irq: irq0 00400000
[ 43.432350] mmcblk0: error -110 transferring data
[ 43.440915] end_request: I/O error, dev mmcblk0, sector 0
[ 43.450091] Buffer I/O error on device mmcblk0, logical block 0

i guess it could be the 20MHz clock not entirely correct. will look into it further.

phew ...

- lt

[touch of jobo]

Booting on my telus diamond touch/x11-initrd it boots to the screen calibration at which point clicking the xhairs has no effect and the device seems stalled.
I have the same results. This calibration screen doesn't do anything for me. You still have to go through the 'normal' (msm_ts.c) calibration during boot (or later).
[...] The screen also has a bit of a greenish/blueish tint to it. [...] though the colors are messed ofc.
I would say they're even more messed up.
http://forum.xda-developers.com/attachment.php?attachmentid=137369&thumb=1
I hoped it wouldn't do that.. :( ..thanks for testing :)

-Jobo

[NetRipper]

i had to use my best guess to crunch the A0/A4 numbers and then out of sudden saw a big step. if the debug log didnt lie to me, the diamond seems to have the moviNAND on SDC2. i got the device /dev/mmcblk0 created and part of the MMC commands ran successfully. although if i "dd if=/dev/mmcblk0" it still times out.
You really make me anxious for a .diff. I'd love to test if it also detects something on my Raphael (even if it is inaccessible).

@touch of jobo, your latest screenshots show virtual keyboard somewhere in the middle. Did you make it draggable? Do you have some time to make a clean diff, possibly without the 'landscape keyboard' (as that's not compatible with diamond). We can commit that to git then, would be a nice addition.

[Untouchab1e]

hallo to everybody,
i am one of many people who want to get android 1.0 running on the diamond and i want to develop with you but i don`t know why android doesn´t work and how progressed you are so far so please can someone post a short "diary of android for diamond"


i have read in archives that we were not able to boot android because we do not got a SD card is that true and if it is so why we do not rename in the registry the attributs of internal storage as we want to?


sorry for my bad english i am a kurd


If you had read the first post in this topic, you would have found this (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux) which is what you are looking for..


EDIIT: lol.. I suddenly have no idea where i found that message.. wrong page

[touch of jobo]

@touch of jobo, your latest screenshots show virtual keyboard somewhere in the middle. Did you make it draggable? Do you have some time to make a clean diff, possibly without the 'landscape keyboard' (as that's not compatible with diamond). We can commit that to git then, would be a nice addition.

Yes, I made it draggable, because it would sometimes get in the way. First I made it toggleable between it eating clicks and clicks going right through it, but this seemed more useful.. It still eats clicks. To change that, make msm_ts_handler always return 0 or ignore the return value in msm_ts.c:do_softint(). It is now also completely flicker free (even through transition animations or dragging the curtain down/up) by calling msm_ts_handler(0,0,-1) right before starting dma. When dragging the vkeyb around, it leaves a trail of keyboards (http://members.home.nl/j0b0/vkeyb-drag.jpg), because I don't know how to 'undraw'.

Here is a git-diff (http://members.home.nl/j0b0/linux-htc-msm-2.6.25-jobo20081215.diff.txt). It is pretty clean, as I have updated to the latest git last weekend.
I just changed MSM_VKEYB_LEFT and _TOP from macros to static globals, keeping them the same names and in the same spot, because this keeps the diff smaller. It would probably be more ..uhm.. appropriate to lose the MSM_ and make them lower case and move them with the other static globals, but that would give a bunch of small modifications all over the place.
(There are some differences where I've been playing with color spaces, which doesn't have the desired effect anyway (as I'm probably barking up the wrong tree on that :( ) but hey.. Just ignore those.)

Does the landscape thing break the kernel on Diamond, or does it just not do anything? I left that code in there, but #if 0'd it out.

-Jobo

[liamhere]

if anyone has a problem with this please pm me

will remove thanks posts....and other pointless posts

hope this will then put the thread back on track :)

[etiger13]

if anyone has a problem with this please pm me

will remove thanks posts....and other pointless posts

hope this will then put the thread back on track :)

can you do that with the Vogue thread too?

[dcordes]

liamhere: wood welcome some moderation on the kaiser thread as well. thanks a lot. http://forum.xda-developers.com/showthread.php?t=396782&page=178

[touch of jobo]

No, my colors do not look normal with the git kernel, but your changes DO make it look normal!

Can you confirm whether this only affects colors for the tux logo, the console, probably even X, but not android? If it looks good everywhere, I'm now officially chasing a ghost.
With the 444 kernel that is, not the mdptest. Using different dma config flags makes things worse, unless this was just not the right combination of flags.

Setting the offsets and lengths of fb_info.fb_var_screeninfo.red, .green and .blue does affect how the tux blits and how the X server draws, but not androids renderer. Not to the point where the output _looks_ different anyway.

Now it appears that android does at least know about the RGB packing settings. In its log, it says:

With the original RGB565 settings:
I/SurfaceFlinger( 291): Client API: OpenGL ES
I/EGLDisplaySurface( 291): using (fd=18)
I/EGLDisplaySurface( 291): id = msmfb
I/EGLDisplaySurface( 291): xres = 480 px
I/EGLDisplaySurface( 291): yres = 640 px
I/EGLDisplaySurface( 291): xres_virtual = 480 px
I/EGLDisplaySurface( 291): yres_virtual = 1280 px
I/EGLDisplaySurface( 291): bpp = 16
I/EGLDisplaySurface( 291): r = 11:5
I/EGLDisplaySurface( 291): g = 5:6
I/EGLDisplaySurface( 291): b = 0:5
I/EGLDisplaySurface( 291): width = 51 mm (239.058823 dpi)
I/EGLDisplaySurface( 291): height = 38 mm (427.789459 dpi)
I/EGLDisplaySurface( 291): refresh rate = 60.00 Hz
I/SurfaceFlinger( 291): OpenGL informations:
I/SurfaceFlinger( 291): vendor : Google Inc.
I/SurfaceFlinger( 291): renderer : Android PixelFlinger 1.0
I/SurfaceFlinger( 291): version : OpenGL ES-CM 1.0


With the RGB444 xxRrrrGg-ggxxBbbb packing:
I/SurfaceFlinger( 291): Client API: OpenGL ES
I/EGLDisplaySurface( 291): using (fd=18)
I/EGLDisplaySurface( 291): id = msmfb
I/EGLDisplaySurface( 291): xres = 480 px
I/EGLDisplaySurface( 291): yres = 640 px
I/EGLDisplaySurface( 291): xres_virtual = 480 px
I/EGLDisplaySurface( 291): yres_virtual = 1280 px
I/EGLDisplaySurface( 291): bpp = 16
I/EGLDisplaySurface( 291): r = 10:4
I/EGLDisplaySurface( 291): g = 6:4
I/EGLDisplaySurface( 291): b = 0:4
I/EGLDisplaySurface( 291): width = 51 mm (239.058823 dpi)
I/EGLDisplaySurface( 291): height = 38 mm (427.789459 dpi)
I/EGLDisplaySurface( 291): refresh rate = 60.00 Hz
I/SurfaceFlinger( 291): OpenGL informations:
I/SurfaceFlinger( 291): vendor : Google Inc.
I/SurfaceFlinger( 291): renderer : Android PixelFlinger 1.0
I/SurfaceFlinger( 291): version : OpenGL ES-CM 1.0

..yet when I draw 565 data in the frame buffer it looks normal in both cases, but also it makes no difference to how android draws. :confused: The only thing so far that does affect how android looks is the mdp dma flags but that just looks too ugly. Could it be an android opengl thing rather than a kernel thing?
Will continue to 'investigate' (tinker around)..

-Jobo


BTW: Changing those color settings has weird side effects:
The X server doesn't stay up after boot, like it would with normal settings,
Killing a manually started X server does drop back to or redraw the console, like it would with normal settings,
The server crashes as soon as I kill 1 client, with the normal settings it would stay up.

[dcordes]

Jobo, I'm excited X works that well on the raphael/diamond. your observations seem to exclude the refresh problem we have on the kaiser and polaris. need to sort out what is the difference between the two framebuffer codes.

Angstrom uses tslib for the ts input. If you want to mess with the touchscreen in X in the angstrom images, check /etc/ts.conf /etc/profile.d/* . in the conf file play around with the settings - I have no clue what you need for raph ts driver. make sure you set the correct input device env var in the profile.d file.
You can calibrate (produce a /etc/pointercal file) when you have the ts-calibrate. try opkg update; opkg install tslib-calibrate . you could use nfs.
If you need anything special in the image, tell me what and I can produce a few builds. btw everybody is welcome to ask for such images. If you want to roll your own,I detailed on openembedded a few pages earlier.

edit: X init is either /etc/init.d/xserver-nodm or /etc/init.d/gpe-dm

btw venus100 (xperia x1) has the same color problem like raph800, see http://forum.xda-developers.com/showthread.php?t=456595

[lavender.t]

You really make me anxious for a .diff. I'd love to test if it also detects something on my Raphael (even if it is inaccessible).


alrighty as you wish ... :) there is a new file clock-7x00.c to be added to the mach-msm/ dir. the mmc-diff.patch is my git index diff (some unrelated files removed), and config.diff are 2 extra debug switches i enabled in the .config file.

<DISCLAIMER: blahblah ... :D>

- lt

[NetRipper]

Does the landscape thing break the kernel on Diamond, or does it just not do anything? I left that code in there, but #if 0'd it out.
Thanks for the .diff. I just took a quick look, and looks good. I'm going to take a better look tomorrow... I'll clean out some of the color stuff though. It can be added when there's a working setup for both RAPH800 and RAPH100 (and equally for DIAM*).

To answer your question, the keyboard bit is always '1' on Diamond. So when the kernel boots, it always detects it as being landscape, making the keyboard vertical right away from boot. Makes it a bit annoying for Diamond ;) Would be nice if there's some runtime way to detect if it's a Diamond or a Raphael. Anyway, somewhere in the near future we may need to split Raphael and Diamond anyway, giving the Diamond its own board file and MTYPE.

@lavender.t, thanks for the diff. I'm aware of the disclaimer :p I'm also going to take a look at that tomorrow.

Need sleep.

[touch of jobo]

[...] I'll clean out some of the color stuff though. It can be added when there's a working setup for both RAPH800 and RAPH100 (and equally for DIAM*). [...]
Sure. I would suggest to only pick out the keyboard drag and non-flicker stuff. The colorspace stuff hasn't given any real success yet (not in android anyway).
-Jobo

[maejrep]

Can you confirm whether this only affects colors for the tux logo, the console, probably even X, but not android? If it looks good everywhere, I'm now officially chasing a ghost.
With the 444 kernel that is, not the mdptest. Using different dma config flags makes things worse, unless this was just not the right combination of flags. I'd love to, but right now I can't get my phone to boot beyond the rtc clocks message, followed by 4 trout_pwrsink_set lines. As of that point, I can only see the tux logo in color, but that logo is correct. If I can figure out why I don't get to the calibration, then I can try to find out the rest of what you're asking.

[evilxhwnd]

I'd love to, but right now I can't get my phone to boot beyond the rtc clocks message, followed by 4 trout_pwrsink_set lines. As of that point, I can only see the tux logo in color, but that logo is correct. If I can figure out why I don't get to the calibration, then I can try to find out the rest of what you're asking.

I am having the same issue on my fuze.

[maejrep]

I'd love to, but right now I can't get my phone to boot beyond the rtc clocks message, followed by 4 trout_pwrsink_set lines. As of that point, I can only see the tux logo in color, but that logo is correct. If I can figure out why I don't get to the calibration, then I can try to find out the rest of what you're asking.

So, with tmzt_'s help on irc, I stepped through each revision starting with the one known-good revision that I had tried in the past, and found the usb-ether revision to be the offending one that made my phone stop booting. The frustratingly easy fix was to disable USB_FUNCTION_ETHER in .config :eek:

Anyway, I'm able to boot again -- anyone with raph800 having that problem, you can try disabling that, or I'm gonna try another fix now. When I plug USB into my phone, I get a popup asking if I want to connect in activesync mode or in "disk drive" mode. I've been selecting disk drive mode, because it registers as a USB mass storage device, which I can mount and transfer my newly compiled kernel over with. But I think that's part of the problem, since as of when haret takes over, the USB interface might still be in mass storage mode, which the ether driver isn't able to cope with. I'm gonna try it now with that enabled, and with USB set in activesync mode before booting.

[maejrep]

But I think that's part of the problem, since as of when haret takes over, the USB interface might still be in mass storage mode, which the ether driver isn't able to cope with. I'm gonna try it now with that enabled, and with USB set in activesync mode before booting.
This fixed the problem, btw.

Anyone with a Sprint (or other) raph800 getting the kernel lockups after trout_pwrsink_set lines, make sure you set the phone back to "activesync" mode before starting up haret.

# ping 192.168.2.206
PING 192.168.2.206 (192.168.2.206) 56(84) bytes of data.
64 bytes from 192.168.2.206: icmp_seq=1 ttl=64 time=2.57 ms
Success :)

Jobo, now I think I'll be able to try out android to see how the colors are affected, since this kernel also has your 444 fix for msm_fb.

NetRipper: you should choose a not-so-standard subnet for the phone's IP next time :p my router uses 192.168.0.0 so I had to change the IP on the phone for it to work.

[NetRipper]

NetRipper: you should choose a not-so-standard subnet for the phone's IP next time my router uses 192.168.0.0 so I had to change the IP on the phone for it to work.
Heh, ok. I'll think about that next time. Good job tackling the boot problem. Updated wiki faq (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux#FAQ) with the info.

[touch of jobo]

[...] yet when I draw 565 data in the frame buffer it looks normal in both cases, but also it makes no difference to how android draws. :confused: The only thing so far that does affect how android looks is the mdp dma flags but that just looks too ugly. Could it be an android opengl thing rather than a kernel thing? [...]

I am now convinced that my earlier hack to draw 444 into the fb fixes everything except android, because android just draws 565 no matter what you tell it.
Color space conversion in android is done in at least a couple of places (opengl/libagl, surfaceflinger/blurfilter, pixelflinger/scanline, ...) but I'm not sure whether it is viable to go in there.. ..if anyone has a setup in place that allows experimenting with that, I would be interested to hear test results.

Does anyone know whether the opengl and related libs in android 1.0 are the same as in 0.8 or whether I can replace the libs in the test initrd with those from 1.0?

Another approach would be that the hardware is capable of working with 565 pixel data but its set to the wrong mode. This must happen during kernel boot as I can't imagine that windows would run in 444 if it could run in 565 mode.

-Jobo

[maejrep]

Another approach would be that the hardware is capable of working with 565 pixel data but its set to the wrong mode. This must happen during kernel boot as I can't imagine that windows would run in 444 if it could run in 565 mode.

-Jobo

Using this as inspiration, I dumped the framebuffer memory in haret and used fbgrab with 16 and 15 bpp (doesn't support 12bpp I guess?) to convert it to a PNG, and while I'm not sure why only a portion of the image is visible -- I probably screwed something up somewhere :) -- the difference is quite clear between the 16 (565) and 15 (555) bpp images, and seems to indicate that WM is in fact drawing 565 to the framebuffer.

So if I'm doing things correctly, that should mean that either a) WM is capable of putting the display into a 565 mode that we aren't doing in linux, or b) WM is drawing 565 to the framebuffer, and there's a conversion taking place for every frame draw from 565 to 444. Obviously (b) would be incredibly inefficient and it would make more sense for them to just draw 444 to the framebuffer directly, which makes (a) the more plausible reason. Does my reasoning seem sound?

If I knew how to dump the correct area of memory to see capabilities, it might help explain which of the two it is. Any tips or suggestions?

[maejrep]

I am now convinced that my earlier hack to draw 444 into the fb fixes everything except android, because android just draws 565 no matter what you tell it.
Color space conversion in android is done in at least a couple of places (opengl/libagl, surfaceflinger/blurfilter, pixelflinger/scanline, ...) but I'm not sure whether it is viable to go in there.. ..if anyone has a setup in place that allows experimenting with that, I would be interested to hear test results.

-JoboYou're right. I tried android.bin, and saw the same thing that hillsdale did (pictured from the wiki). If you're asking for someone to test your changes, let me know -- if you're asking if I've done it myself, nevermind :)

[computerguru]

I have a question about how the display resolution works with the kernel. Is it a) using the video mode from Windows Mobile, or b) does the kernel set the mode during the boot stage?

If it's (a), I could try writing some code to access the GPE driver from within Windows Mobile to change the color mode (should be able to do it via GetModeInfo and SetMode...if the driver allows it.)

If it's (b), what happens if we skip the kernel-based video mode initialization and use the mode provided by Windows Mobile?...since it may already be in the 565 mapping. If you want, I could write some code to do some visual comparisons between different shades of green...to see if Windows Mobile is doing the conversion. The chipset for the DIAM500 should be able to support it...I mean...the TFT supports 262,144 colors.

http://developer.sprint.com/device/loadDevice.do?id=859d2565-15e1-47ca-aba0-3fcb4d5bc3b2

Sorry if my questions have already been answered...but I don't have time at the moment catch up from page 80 onwards...lol. :)

[kurgandk]

hah!!

I love the update subject line... :D

[lavender.t]

finally got a clock and timing combination that works for my diamond ! was able to mount my "internal storage":

$ telnet 192.168.0.206
Trying 192.168.0.206...
Connected to 192.168.0.206.
Escape character is '^]'.



BusyBox v1.13.0 (2008-11-28 00:52:00 PST) built-in shell (ash)
Enter 'help' for a list of built-in commands.

# dd if=/dev/mmcblk0 count=1 | hd
1+0 records in
1+0 records out
00000000 eb fe 90 4d 53 57 49 4e 34 2e 31 00 02 08 20 00 |...MSWIN4.1... .|
00000010 02 00 00 00 00 f8 00 00 00 00 00 00 00 00 00 00 |................|
00000020 00 00 78 00 00 1e 00 00 00 00 00 00 02 00 00 00 |..x.............|
00000030 01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |................|
00000040 80 00 29 12 00 43 08 20 20 20 20 20 20 20 20 20 |..)..C. |
00000050 20 20 46 41 54 33 32 20 20 20 00 00 00 00 00 00 | FAT32 ......|
00000060 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |................|
*
000001f0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 55 aa |..............U.|
00000200
# mkdir mnt
# mount /dev/mmcblk0 /mnt
# ls /mnt
Content/ MSMETADATA/ WMPInfo.xml* android/
EncFiltLog.menc* My Documents/ YouTube/ storage/
Live Search/ VerInfo.txt* album.vol* wmdrm/
# ls /mnt/android/
android-initrd.cpio.gz* haret-big.exe*
default.txt* haret.exe*
default1.txt* initrd-android-raphael.cpio.gz*
default2.txt* initrd-netripper-busybox-usb.cpio.gz*
default3.txt* initrd.cpio.gz*
defaulta.txt* ramdisk.img.gz*
defaulto.txt* zImage*
haret-0.5.2.exe* zImage.old*
#

i'll give a few more tries after i wake up to see how stable it is.

cheers ! :D

- lt

[beboxed]

WOOOT! Go lavender.t!

btw, is this in DMA mode or PIO?

[rl2000]

Nice dude! Nice!

[NetRipper]

finally got a clock and timing combination that works for my diamond ! was able to mount my "internal storage":
Awsome. Good job. :)

I'll await a diff ;) Good night for now, well deserved ;)

[balsat]

finally got a clock and timing combination that works for my diamond ! was able to mount my "internal storage":


I love you man!!! ;)
Can't wait for a diff.

[nuke1]

Not sure if this is relevant, or even if it's known:

http://www.pocketnow.com/index.php?a=portal_detail&t=news&id=6459

[liamhere]

hah!!

I love the update subject line... :D


glad you liked it.....and the thread is running good again :)

[holydhaliwal]

finally got a clock and timing combination that works for my diamond ! was able to mount my "internal storage":
- lt

Does this mean that we will soon be able to load up android 1.0? or am i missreading?

[lavender.t]

if you cant wait to try, here are the diff file and clock-7x00.c. it is not final yet so forgive the code quality. and ... yes i realized the msm_sdcc.c change can be done by the kernel parameters. :)

- lt

[Willis111]

if you cant wait to try, here are the diff file and clock-7x00.c. it is not final yet so forgive the code quality. and ... yes i realized the msm_sdcc.c change can be done by the kernel parameters. :)

- lt

I'm getting a warm fuzy feeling in my pants right now.

[maejrep]

if you cant wait to try, here are the diff file and clock-7x00.c. it is not final yet so forgive the code quality. and ... yes i realized the msm_sdcc.c change can be done by the kernel parameters. :)

- lt

Any tips on what I should look for to see if raph800 is different, and/or how to debug it? Right now it boots with your patch, but there's no /dev/mmcblk*
# cat /sys/kernel/debug/mmc0/ios
clock: 0 Hz
vdd: 0 (invalid)
bus mode: 1 (open drain)
chip select: 0 (don't care)
power mode: 0 (off)
bus width: 0 (1 bits)
timing spec: 0 (legacy)

# dmesg | grep -i slot
[ 1.510183] mmc0: Slot eject status = 0


[ 1.490183] mmc0: Qualcomm MSM SDCC at 0x00000000e1001000 irq 26,0 dma 8
[ 1.500427] mmc0: 4 bit data mode enabled
[ 1.508026] mmc0: MMC clock 2640000 -> 18480000 Hz, PCLK 66000000 Hz
[ 1.510183] mmc0: Slot eject status = 0
[ 1.520061] mmc0: Power save feature enable = 0
[ 1.527598] mmc0: DM non-cached buffer at ffc03000, dma_addr 0x15bc9000
[ 1.530335] mmc0: DM cmd busaddr 364679168, cmdptr busaddr 364679936
[ 1.540122] mmc0: Polling status mode enabled

[tughi]

I just compiled the cupcake branch... In the emulator you can use now the on-screen keyboard :)

Update: screen shot

http://img149.imageshack.us/img149/1076/keyboardbm0.th.png (http://img149.imageshack.us/my.php?image=keyboardbm0.png)

[holydhaliwal]

I just compiled the cupcake branch... In the emulator you can use now the on-screen keyboard :)

Awesome, would you be able to post a screenshot of the keyboard? i can't locate one online for some reason...

[lattiS710]

Hi guys, I dont have time to read all this 114 pages and I am wondering, how does this project goes? Are we going to be able to have Linus/Android on our diamonds soon?

peace

[jeforce]

Hi guys, I dont have time to read all this 114 pages and I am wondering, how does this project goes? Are we going to be able to have Linus/Android on our diamonds soon?

peace

Look at the wiki => http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux ;)

[Black_Angel]

When will somebody update the wiki with the latest version of Android adapted for our HTC Touch Diamond?

Because it seems the on-screen keyboard is functionnaly etc...

[brumle80]

When will somebody update the wiki with the latest version of Android adapted for our HTC Touch Diamond?

Because it seems the on-screen keyboard is functionnaly etc...

less than 24 hours since the post about that, hint hint....

[liamhere]

Hi guys, I dont have time to read all this 114 pages and I am wondering, how does this project goes? Are we going to be able to have Linus/Android on our diamonds soon?

peace

as title says...." not yet working"....so to keep upto date.. if you don't have time to read all those pages .....please try to read at least the last page....you may get a hint of where the project stands at this moment in time.

[dcordes]

sorry for adding even more noise on top of the most important mmc topic.

can someboy please try reading and writing from /dev/smd0 on the raph100 raph800 and diam? make sure the radio is turned on in windows mobile and ready to make calls at linux boot time. tia

[Willis111]

When will somebody update the wiki with the latest version of Android adapted for our HTC Touch Diamond?

Because it seems the on-screen keyboard is functionnaly etc...


sorry for adding even more noise on top of the most important mmc topic.


most important mmc topic? screw that someone just got the on-screen keyboard working! Read the wiki, the latest version of android ported to the touch diamond is available there. If you actually read the post that ? made about compiling cupcake and booting it with an EMU ... you'd realise that he didn't boot it using a phone, he did it all on a computer which would make that the latest version of android ported to the touch diamond emulator. I think the holdup is that though there has been a huge breakthrough by my personal hero lavender.t in mounting his sd card in android no one else has been able to duplicate it yet and we're lacking storage space to run android 1.0. If I'm wrong here feel free to correct.

[sibbor]

I just compiled the cupcake branch...Looks great! Hopefully we'll be able to progress further as soon as we get the mmc stuff working out. l.t did a really great job there :D

As most of us here already knows "cupcake" is based on the Linux kernel 2.6.27.

Just to keep the spirit high I'm quoting what they say about the "cupcake" development branch (http://source.android.com/roadmap/cupcake):

To check out the cupcake branch:
mkdir cupcake # create a new client directory
cd cupcake
repo init -u git://android.git.kernel.org/platform/manifest.git -b cupcake
repo sync

Applications

* MMS
o New features
+ Save attachments from MMS.
o Significant bug fixes
+ Faster conversation list scrolling

* Email
o Significant bug fixes
+ Accounts that were marked "never check" are not auto-checked.
+ Date & time displayed using user preference (e.g. 24 hr vs. AM/PM).
+ cc: displayed in message view.
+ Relaxed POP3 parser rules so it works with non-compliant email servers.
+ Password quoting bugs in IMAP. Makes it work for users with funny chars in their password (e.g. spaces).
+ Various sources of errors in auto & manual account setup.
+ Improvements on how we report various connection errors. Makes it much easier for user to diagnose failed account setups.
+ New-mail notifications for POP3 accounts.
+ Properly recover from POP3 connection failures, so that the next connection has a chance of working properly.
+ Remove automatic accounts setup entries that were broken or not testable. Minor fixes to a few of the remaining entries. Improvements to warning dialogs used for a few special cases.
+ New accounts are now set to check every 15 minutes (instead of defaulting to "never").
+ Fixed a bug causing approximately 1 in 25 outbound messages to freeze up the IMAP connection (to a Gmail based server) when transferred to the Sent folder. This broke the entire connection so new messages could not be downloaded either.
+ Unit test framework so Email can be extended & tested more reliably.
+ Fix IMAP manually-created accounts so message delete works properly.

* Alarm Clock
o Significant bug fixes
+ Alert now plays audio/vibe directly, rather than through AlarmManager. AlarmClock alert starts playing audio/vibe in its IntentReceiver, rather than on activity start. These changes should prevent alarms from being blocked by modal dialogs.

* Package Installer
o Significant bug fixes
+ Bugs related to replacing existing applications.

* Settings
o New features
+ New menu option to list running processes in Settings->ManageApplications.

* Music
o New features
+ Music playback fades in after suspending for phone call.
+ New media search intent allows for 3rd party apps to launch or respond to media searches based on artist, album, or title.
Affects: Music Player, YouTube, Browser applications.

* Browser
o New features
+ Updated WebKit browser core, synced with Nov 2008 WebKit version.
+ Support for new, optimized JavaScript engine (SquirrelFish).
+ Copy / paste is enabled in the browser. To copy with touch, press and hold the shift key and select the text. Releasing the shift key or ending the touch drag copies the text. To copy with the trackball, press and hold the shift key, move the cursor to the selection start, click the trackball, and move the trackball to the extend the selection. Releasing the shift key, or clicking the trackball a second time, copies the text.
+ Find is enabled in the browser. To find text, choose it from the menu and type the text to find.
+ Drawing has been sped up substantially by supporting partial content invalidates and partial screen invalidates. Pages with animations are 5x faster.

* VoiceDialer
o New features
+ VoiceDialer supports 'open app' command

* Camera/Gallery
o New features
+ Video recorder mode
+ Share intent for videos
+ Video thumbnails
+ Local file playback

Download manager

* New features
o Support for HTTP codes 301, 302, 303 and 307 (redirects).
o HTTP code 503 is now handled, with support for retry-after in delay-seconds.
o Downloads that were cleanly interrupted are now resumed instead of failing.
o Applications can now pause their downloads.
o Retry delays are now randomized.
o Connectivity is now checked on all interfaces.
o Downloads with invalid characters in file name can now be saved.

Framework

* New features
o Support of touch events in WebView.
o New JavaScript engine (SquirrelFish) in WebView.
o Input method framework, for soft keyboards and other on-screen input methods. Includes new APIs for applications to interact with input methods, and the ability for third party developers to write their own input methods.
o Access to the raw audio data for playback and recording from application code.
o New PendingIntent.FLAG_UPDATE_CURRENT option.
o Support for top-level boolean resources.
o Tactile feedback to the LockPatternView. Tactile feedback can be enabled/disabled by going to Settings > Security & location and then checking/unchecking "Use tactile feedback". Note that this can be used independently of the visual feedback of the lines ("Use visible pattern"). Thus it gives users a middle ground between showing the lines on the screen and having no feedback at all.
o PackageManager changes to support un-installation of partially installed applications. Added new flag PackageManager.GET_UNINSTALLED_PACKAGES to include partially installed apps in all relevant PackageManager api's. ManageApplications screen now lists such partially installed apps and the user can uninstall these applications completely.
o Support third party updates of system applications. New menu options in Settings->ManageApplications to list updated system applications.
o Framework support to list current running processes. New API in ActivityManager.
o Framework feature to declare required configurations by applications. New manifest attribute uses-configuration in android manifest.
o Hardware accelerated video encode (video recorder) in opencore.
o Simplified SREC speech recognition API available.
o Streaming audio I/O for applications.

* Significant bug fixes
o Fixed issues with saving state in the view hierarchy, so that you can properly subclass from something like TextView and create your own state that inherits from that provided by TextView.
o TextView now implements onKeyMultiple(), so that flinging the trackball will result in accelerated scrolling. This required some changes to movement methods, and included some improvements to the acceleration computed when flinging.
o Framework bug fixes in PackageManager to share/un-share permissions for applications with shared uid's.
o Significant rework of Settings->ManageApplications Performance and UI enhancements.
o A number of settings in android.provider.Settings.System were moved to android.provider.Settings.Secure. Only system software can modify these settings. Additionally, a new permission, WRITE_SECURE_SETTINGS, is required to access these settings. The old constants in Settings.System have been deprecated. It is possible to read settings values via Settings.System using the deprecated constants. However, attempts to modify these settings via Settings.System will result in a log message and the setting value will be left unchanged.
o Many bug fixes in the media framework


Bluetooth

* New features
o Support for A2DP & AVRCP profiles.

* Significant bug fixes
o First connection after pairing always fails on many carkits.
o Mini Cooper and some late model BMW cars fail to use Bluetooth or take 2 minutes for Phone Book transfer.

System software

* New features
o New kernel based on Linux 2.6.27.
o Improvements to the wakelock API.
o Work to transition to the USB Gadget Framework underway.
o Basic x86 support.


Radio & Telephony

* New features
o SIM Application Toolkit 1.0.
o Green CALL button is no longer a shortcut for "add a new call". This has been a rarely used feature and confusing if triggered accidentally.
o Longer in-call screen timeout when using the speakerphone.
o "Show dialpad" / "Hide dialpad" item added to the in-call menu, to make it easier to discover the DTMF dialpad.

* Significant bug fixes
o An obscure case where the Phone UI could cause the device to not go to sleep on its own. This would happen if user bails out of the in-call screen by hitting HOME, followed by the call disconnecting remotely.
o Don't allow a single tap to open the in-call dialpad. It is now required to touch and drag it. This makes it much harder to accidentally open the dialpad by touching the screen with your face.

[tughi]

As most of us here already knows "cupcake" is based on the Linux kernel 2.6.27.

Last time i checked cupcake was still based on 2.6.25... At least this is reported by the emulator's "about phone". But I guess this will change soon.

http://img357.imageshack.us/img357/6798/statusaf9.th.png (http://img357.imageshack.us/my.php?image=statusaf9.png)

[dcordes]

how about a userspace thread and keep this one kernel development only?

[touch of jobo]

[...] I think the holdup is that though there has been a huge breakthrough by my personal hero lavender.t in mounting his sd card in android no one else has been able to duplicate it yet and we're lacking storage space to run android 1.0. If I'm wrong here feel free to correct.
well actually...
HaRET will load, unpack and boot into an image just over 33M (!! Thirty-three of your Earth megabytes), which would fit a trimmed down android 1.

(I've been (yet unsuccessfully) trying to put together an android 1 image in a single root ramdisk. It boots the kernel, starts android runtime, then complains about not being able to load frameworks (http://members.home.nl/j0b0/android1-jobo-logcat.txt). This has more to do with me not knowing what I'm doing than with android 1 requiring internal storage.)

So not mounting the internal storage should not necessarily stop you from running android 1. (You just need to know what you're doing :)) But still, with no access to (persistent) storage, anything you do is just for the sake of it.

Mounting the internal storage makes it possible to run the vogue way, which is with a small initrd that mounts the /system and /data filesystems from 2 files on internal storage. I'm guessing we need to mount at least the data fs read-write for stuff to work.

[...] can someboy please try reading and writing from /dev/smd0 on the raph100 raph800 and diam? [...]
This boy can. (Sorry ;)) It says:
+CREG: 1,"0460","415B"
+CGREG: 0
@HTCCSQ: 3
+CREG: 1,"11B5","2F2D"
@HTCCSQ: 4
@HTCCSQ: 3
+CREG: 1,"11B5","2F2B"
@HTCCSQ: 4
[... I call my raphael from a landline. The Xs are really numbers .. mine]
+CRING: VOICE
+CLIP: "+315XXXXX15",145,,,""
+CRING: VOICE
+CLIP: "+315XXXXX15",145,,,""
+CREG: 1,"11B5","2F2B"
+CGREG: 0
+CRING: VOICE
+CLIP: "+315XXXXX15",145,,,""
@HTCCSQ: 4
[... I hung up. It goes back to reporting CREGs and HTCCSQs]

..on a GSM raphael. Should I also test with wifi / gps / fm tuner enabled, or is that different radios? What would you like me to write to it?

how about a userspace thread and keep this one kernel development only?
..and change the title to 'Linux on HTC Diamond & Raphael working to some degree. Android? What android?'

-Jobo

[balsat]

I manage to boot into a ext2 partition on /dev/mmcblkp and got android 1.0 running on it... I unpacked http://www.netripper.com/raphael/20081204-01_raph_diam_angstrom_usb/initrd-angstrom-usb.cpio.gz and unpacked data and system from Android image to the sdcard.

But after i rebooted there is problems starting Android... strange.

[dcordes]

This boy can. (Sorry ;)) It says:
+CREG: 1,"0460","415B"
+CGREG: 0
@HTCCSQ: 3
+CREG: 1,"11B5","2F2D"
@HTCCSQ: 4
@HTCCSQ: 3
+CREG: 1,"11B5","2F2B"
@HTCCSQ: 4
[... I call my raphael from a landline. The Xs are really numbers .. mine]
+CRING: VOICE
+CLIP: "+315XXXXX15",145,,,""
+CRING: VOICE
+CLIP: "+315XXXXX15",145,,,""
+CREG: 1,"11B5","2F2B"
+CGREG: 0
+CRING: VOICE
+CLIP: "+315XXXXX15",145,,,""
@HTCCSQ: 4
[... I hung up. It goes back to reporting CREGs and HTCCSQs]

..on a GSM raphael. Should I also test with wifi / gps / fm tuner enabled, or is that different radios? What would you like me to write to it?


echo AT to it while reading in the background. you should get 0. echo ATV1 and you should get OK


..and change the title to 'Linux on HTC Diamond & Raphael working to some degree. Android? What android?'

-Jobo

why not

[lavender.t]

echo AT to it while reading in the background. you should get 0. echo ATV1 and you should get OK


i got similar results as jobo on my diamond when calling in:
# cat /dev/smd0
+CRING: VOICE
+CLIP: "+1425xxxxxxx",145,,,""
+CRING: VOICE
+CLIP: "+1425xxxxxxx",145,,,""
+CRING: VOICE
+CLIP: "+1425xxxxxxx",145,,,""
+CRING: VOICE
+CLIP: "+1425xxxxxxx",145,,,""
+CRING: VOICE
+CLIP: "+1425xxxxxxx",145,,,""
+CRING: VOICE
+CLIP: "+1425xxxxxxx",145,,,""
+CRING: VOICE
+CLIP: "+1425xxxxxxx",145,,,""
@HTCDIS: 0013,1

when writing AT or ATV1 cmd to /dev/smd0, no response was seen from the cat side.

- lt

[lavender.t]

no further news. just cleaned up the code in clock-7x00.c (now support setting clocks for SDC 1 - 4), pulled in the rest of the clock code from clock.c. also temporarily added a kernel option "msmsdcc_id=" (1 - 4, my DIAM100 uses 2) to allow changing the sdc to try without recompiling the kernel.

the diff is attached. i havent seen any one else reported success or failure in diamond yet. i hope this patch could at least workaround diamond's mmc issue.

maejrep and netripper found the patch not working for raphael. i didnt have a clue yet but it appeared to be the clock still not correct for raphael.

i may try if android 1 works in my diamond when i got some time to spare. there are still quite some other significant issues to tackle with so we still need to use some patience ... :(

happy holidays, at least ... :)

- lt

[dcordes]

ok so obviously AT tx doesn't work. we should fix this in order to talk to the modem from linux. I don't know where the rx/tx offsets from smem base are set in this tree.

lt, did you try adding the clock rate calculation thing cr2 proposed?

[touch of jobo]

Does anyone know how to read / interpret the table in /drivers/video/msm/mdp_csc_table.h or for that matter the mdp_default_css[] at the top of mdp.c ?
Thanks,
-Jobo

[abc12345]

I manage to boot into a ext2 partition on /dev/mmcblkp and got android 1.0 running on it... I unpacked http://www.netripper.com/raphael/20081204-01_raph_diam_angstrom_usb/initrd-angstrom-usb.cpio.gz and unpacked data and system from Android image to the sdcard.

But after i rebooted there is problems starting Android... strange.

does this mean that you have partial sd support?

[mdrobnak]

I've read through some of this thread, and I'm happy to see things are moving along. I've got an HTC Fuze. I'd like to know: what can I do to further along development? I'm rusty in C/C++, and have done assembly before on x86...

A few questions:
Are we reverse-engineering from the windows drivers? Are we allowed to do that?
How have we gotten as far as we have in terms of memory locations, IRQs, etc?

Basically, what kind of tools and refrences should I be looking at to try and get this storage & audio stuff (what seems to be the next major stumbling blocks) going...

Thanks,
-Matt

[maejrep]

echo AT to it while reading in the background. you should get 0. echo ATV1 and you should get OK

Does not work on raph800:
root@htcraphael:~# cat /dev/smd0
cat: can't open '/dev/smd0': No such device
root@htcraphael:~# echo "AT" > /dev/smd0
-sh: cannot create /dev/smd0: No such device
[ 4.627507] smd_init()
[ 4.627568] smd_core_init()
[ 4.627659] smd_core_init() done
[ 4.627812] smd_alloc_channel() 'SMD_RPCCALL' cid=2, shared=e0110148
[ 4.628117] smd_alloc_channel() 'SMD_CONTROL' cid=5, shared=e010c120
[ 4.628361] smd_alloc_channel() 'SMD_DATA5' cid=11, shared=e0118198
[ 4.628636] smd_alloc_channel() 'SMD_DATA6' cid=12, shared=e011c1c0
[ 4.628850] smd_alloc_channel() 'SMD_DATA7' cid=13, shared=e01201e8
[ 4.629094] smd_alloc_channel() 'SMD_CS_A2M' cid=33, shared=e0114170

Also seeing:# echo 1 > /sys/module/smd/parameters/debug_mask
...
[ 138.500355] smsm_set_interrupt_info 0 0 -> 2400432 0
[ 138.504384] smsm_set_sleep_duration 971 -> 1300
[ 138.508198] smsm_change_state a29
[ 138.510142] smsm_change_state 2829
[ 138.510142] smsm_change_state 929
[ 138.510142] <SM 00000929 00010d29>
[ 138.510142] <SM 00000929 00010d29>
[ 138.580355] smsm_set_interrupt_info 0 0 -> 2400432 0
[ 138.584231] smsm_set_sleep_duration 1300 -> 972
[ 138.587954] smsm_change_state a29
[ 138.590142] smsm_change_state 2829
[ 138.590142] smsm_change_state 929
[ 138.590142] <SM 00000929 00010d29>
[ 138.590142] <SM 00000929 00010d29> If that helps any :x

[balsat]

does this mean that you have partial sd support?

I split my sdcard in 2, the first is a fat partition and the second is a ext3 for Linux.
The I unpacked the anstrom initrd on the ext3 partition, changed my default.txt for haret.exe to

set MTYPE 1910
set KERNEL zimage
set CMDLINE "root=/dev/mmcblk0p2 rootdelay=15 mem=96"
set RAMSIZE 0x6000000
set RAMADDR 0x100000000
bootlinux

[Untouchab1e]

And then what happened?

[balsat]

It boot up on a Linux shell, but i can't start android now because of a missing binderdev!

Merry Christmas

[lavender.t]

lt, did you try adding the clock rate calculation thing cr2 proposed?

sorry for being late, dcordes. blame the holiday ... :)

i was really excited when i saw cr2's formula. but when i looked closer i had the same question as netripper that there was no where to derive the real freq or divisor value from those parameters. also tried the formula for 144kHz parameters but i found the A4 value was different from what i got from the code. i thought 3 shift values could be wrong (missed by 1).

i sent message to cr2 and hasnt been answered yet. :(

- lt

[dcordes]

sorry for being late, dcordes. blame the holiday ... :)

i was really excited when i saw cr2's formula. but when i looked closer i had the same question as netripper that there was no where to derive the real freq or divisor value from those parameters. also tried the formula for 144kHz parameters but i found the A4 value was different from what i got from the code. i thought 3 shift values could be wrong (missed by 1).

i sent message to cr2 and hasnt been answered yet. :(

- lt

Heya, I think it's better to drop a line in the channel. he frequently reads the logs.

[abc12345]

It boot up on a Linux shell, but i can't start android now because of a missing binderdev!

Merry Christmas

So what do we need to do to fix this?

[hah2110]

So how do I actually install android onto my fuze? The newest, working version... These threads are waayyy to cluttered to provide any real benefit.

[balsat]

The Wiki also says BOOTS: OK... Mine doesn't boot....

which zImage, initrd and haret.exe do you use, what does your default.txt say?

[konsta819]

hah2110:

The point is, that this doesn't work. You can try hard, and waste other people's time in asking for help replicating the steps that have been taken up to the point at which the boot fails. Nothing at all right now will give you what you are looking for - not an android phone, or even something to play with the software for curiosity.

You are clearly not a skilled linux developer, and so as much as you'd like to help, the chances that you can contribute anything are incredibly limited. Right now you're just contributing to the 210 pages of clutter that make it hard for the real developers to do their thang.

Bear in mind the first line of the wiki. If you don't know what you're doing, you may brick your phone.

You were essentially told this in the first answer, which is why people by this point are irritated.

[liamhere]

i said i was watching :p

[liamhere]

So how do I actually install android onto my fuze? The newest, working version... These threads are waayyy to cluttered to provide any real benefit.


read read...and read some more

posts like this is why the thread is "cluttered"

[pest666]

Hi, there
I like many others have my diamond and would really love to get android running on it, I found this thread.. and whatta long list of questions.. whithout answers...

this thread is just a list of 120 pages of disrespect and the wiki is useless...

Im a developper myself and have been working on many projects and seen a lot of them die because of structure and focus problems.
lists grow longer and longer with nothing but complains. no really progress..
this is looking like one of these "projects"

so before giving up, bying an iphone, a mac go to a fancy cafe drinking laté and look like im important........


first,
is there a "real" project somewhere, or is there only this meaningless talk ??
whatabout taking an agile aproach, create a real project somewhere ??
where is the developers listed, who is doing what, what order should we fix what ??
why are there focus on the colors when the phone is useless. without the mic and speakers working?
i would love to see any of you guys go to work in a car without the engine, and even without the tires...
if this is fixed, a lot would start using it, interest will grow and maybe a community will appear..


second,
what is it people want ? is it android on the diamond or is it the penguin logo when booting ??
anyone know a project with android running on windows mobile ?
the diamond is born with windows and orig drivers.. maybe it is there we should focus.

im not a fan of the m$ corp and would love android running on linux. but it will take a while.. so until then maybe some of us could get it running on windows, and get focus on android.. or would this be i direct competition to this "project" ?

can android run on windows at all ?


/Happy New years,
David N.

[skiwithpete]

David N.,

http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux

[sensorama]

David N.,

http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux

this made my day.

:D

[pest666]

Hi, there
I like many others have my diamond and would really love to get android running on it, I found this thread.. and whatta long list of questions.. whithout answers...

this thread is just a list of 120 pages of disrespect and the wiki is useless...

Im a developper myself and have been working on many projects and seen a lot of them die because of structure and focus problems.
lists grow longer and longer with nothing but complains. no really progress..
this is looking like one of these "projects"

so before giving up, bying an iphone, a mac go to a fancy cafe drinking laté and look like im important........


first,
is there a "real" project somewhere, or is there only this meaningless talk ??
whatabout taking an agile aproach, create a real project somewhere ??
where is the developers listed, who is doing what, what order should we fix what ??
why are there focus on the colors when the phone is useless. without the mic and speakers working?
i would love to see any of you guys go to work in a car without the engine, and even without the tires...
if this is fixed, a lot would start using it, interest will grow and maybe a community will appear..


second,
what is it people want ? is it android on the diamond or is it the penguin logo when booting ??
anyone know a project with android running on windows mobile ?
the diamond is born with windows and orig drivers.. maybe it is there we should focus.

im not a fan of the m$ corp and would love android running on linux. but it will take a while.. so until then maybe some of us could get it running on windows, and get focus on android.. or would this be i direct competition to this "project" ?

can android run on windows at all ?


/Happy New years,
David N.

answer myself.. ;)
just read the "what is android"

........

Linux Kernel

Android relies on Linux version 2.6 for core system services such as security, memory management, process management, network stack, and driver model. The kernel also acts as an abstraction layer between the hardware and the rest of the software stack.

[pest666]

David N.,

http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux

Wow.. nice pictures.. love them.. yes....
..my mom can crete a wiki with nice pictures...

i WAS interested in participating.. but. it seems i have to read 120 pages "read the wiki", and then look at some nice pictures, boot up my diamond with a precompiled binary, and have no use of my phone..... Why would a developer be interested participating in a project, with nowhere to start and only grumpy comments.

[ermin15]

David N.,


Wow.. nice pictures.. love them.. yes....
..my mom can crete a wiki with nice pictures...

i WAS interested in participating.. but. it seems i have to read 120 pages "read the wiki", and then look at some nice pictures, boot up my diamond with a precompiled binary, and have no use of my phone..... Why would a developer be interested participating in a project, with nowhere to start and only grumpy comments.

What would you suggest us to do. Maybe a forum? Here are enough non-devs who are willing to donate a bit time to help.

[pest666]

my exp on this project is not that deep i actually trying to do some modifications so im able to compile the kernel on a mac. but im just about to just reboot into linux..

as i undestand it, its a branch on the linuxtogo ??
linuxtogo has a lot of projects running:
http://projects.linuxtogo.org/

maybe register a project there ?
or a trac somewhere.

A wiki/forum is not enough.

[skiwithpete]

David N.,
i WAS interested in participating.. but. it seems i have to read 120 pages "read the wiki", and then look at some nice pictures, boot up my diamond with a precompiled binary, and have no use of my phone..... Why would a developer be interested participating in a project, with nowhere to start and only grumpy comments.

Check IRC Freenode channel HTC-Linux... this is IF you're still interested.

[mdrobnak]

Check IRC Freenode channel HTC-Linux... this is IF you're still interested.

I was actually looking for this exact information..it's not in the wiki, or the front page of xda-dev...

I think I shall add this to the Wiki.

-Matt

EDIT: I take that back, it is in the first line..I feel dumb now. :-)

[pest666]

Is there somewhere yet another wiki maybe.. with a haret/linload/android guide, like the build guide. i have been looking arround and yet again I need to find some post in the middle of a 200 posts thread just to get near to an undestanding of what is needed..

I see netripper supplying a lot of exe's but to build one for the diamond myself. ?
how do "stupid little me" do it without having to hardreset 250 times before getting it right??

the "Linux on Raphael and Diamond" wiki could have a usefull link like Research with a sub page getting started howto.

- prereq
- build the kernel
- howto create the exe with/without linload, loadlin or whatever its called
- etc......

/D

[pest666]

I was actually looking for this exact information..it's not in the wiki, or the front page of xda-dev...

I think I shall add this to the Wiki.

-Matt

EDIT: I take that back, it is in the first line..I feel dumb now. :-)

Dont fell that, when you have to filter one line out, its easy to miss. I did it to


I dont know how updates to the wiki is managed ? is it anarchy or ?
If i want to add some small diamond specific howtos as subpages I just do it or ? i should think not ??

[mdrobnak]

I was actually looking for this exact information..it's not in the wiki, or the front page of xda-dev...

I think I shall add this to the Wiki.

-Matt



Dont fell that, when you have to filter one line out, its easy to miss. I did it to


I dont know how updates to the wiki is managed ? is it anarchy or ?
If i want to add some small diamond specific howtos as subpages I just do it or ? i should think not ??

A Wiki is designed to be updated by it's users - I bolded the information for IRC, for instance. You can add new pages as you feel fit...

-Matt

[skiwithpete]

this made my day.

:D

That made my day

:)

[timmy1980]

I am following this thread but do not really have the know how to make any real contribution. All credit to those who do. I just noticed this and wondered if it may be of any help. More than likely I am barking up the wrong tree but just thought I'd run it past those who DO know what they are talking about!

http://forum.xda-developers.com/showthread.php?t=463779

[abc12345]

is the structure the same as in other htc roms?

[amandeepphull]

I've downloaded the file (20081204-01_raph_diam_android_v0.8) and placed it in the Internal Storage (The 4GB storage part), Used the File Explorer to find the file. Launched the file, but it doesent load!

How long am I expected to wait for the kernel to load etc.?

Thanks!

[Willis111]

I've downloaded the file (20081204-01_raph_diam_android_v0.8) and placed it in the Internal Storage (The 4GB storage part), Used the File Explorer to find the file. Launched the file, but it doesent load!

How long am I expected to wait for the kernel to load etc.?

Thanks!

So, if I go by what you wrote you open file explorer, run haret and then just wait. Try hitting the run button and if that generates any errors post them here.

Edit: also there was a problem/solution posted a few pages back. Make sure your device is setup for active sync.

[touch of jobo]

Is this useful?
http://www.phonewreck.com/wiki/index.php?title=HTC_Touch_Pro

-Jobo

[amandeepphull]

Cheers Willis, I shall look into this tomorrow, I have a Massive Financial accounts project in and wanted to test android to take time out :D lol. I will report back tomorrow hopefully!
My Diamond was not set up for activesync at the time (no connection or Sync Partnerships).

When you say Set up for activesync, do you mean I have to execute HaRET while activesync is connected or do you mean, i simply have to make an activesync connection with my laptop, then i can disconnect and try Android?
Regards,
Aman

Quick Edit: My error is:
NET: Registered to protocol family 1
NET: Registered to protocol family 17
clock_late_init() disabled 17 unused clocks
drivers/rtc/hctosys.c: unable to open rtc dev
ice (rtc0) <- I know it's device (rtc0), but i just typed it as it appreared

trout_pwrsink_set:sTUB!
trout_pwrsink_set:sTUB!
trout_pwrsink_set:sTUB!
trout_pwrsink_set:sTUB!

END - no further commands came up.

I'll do some more reading on this tomorrow!


EDIT again: I managed to get it to work by connecting to activesync and following on screen instructions :D

[lavender.t]

spent some time in the holiday and finally got android 1.0 running in my diamond. not really a usable system 'cause as you may have already found out by reading the thread or working on porting that many functions in the kernel still need to be fixed. just did this as a proof of concept -- yes we are getting closer and yes you still need to use patience. :)

some notes on pulling in android 1.0:

1. diamond's built-in mmc cannot be partitioned (at this time). it is not because of linux btw. the problem is WM will not be able to recognize the partitioned mmc in diamond. and unfortunately i still need to use my WM phone before android is fully ready. so i built the system.img and userdata.img as ext2 fs and mount them via loopback devices.

i also tried enabling the mtd + yaffs2 support in the kernel and loopback the ready-made android images. however i was having trouble making the loopback work with yaffs2 and didnt have time to track down. so just took the easy path to use ext2 instead.

2. if we start from our own "init" (for sake of debugging) instead of android's "init", the busybox mdev cannot be used. android "init" also takes care of /dev file system (as well as /proc /sys, etc). running mdev will have confliction and the java vm cannot be launched successfully. without mdev, we have to create a minimal set of devices first (tty, pts, loop, mmcblk, ...) statically to run our telnetd and mount the images.

3. make sure the /data image is big enuf. in the beginning i used only 8MB and the java vm just couldnt launch. the error messages all looked to be corrupted jar files. took me quite some troubleshooting time to find out it was indeed disk full and enlarged the image to 64M. also make sure the tree under /data has permission 0771 'cause files will be created there. on the other hand the /system can be mounted RO.

4. logcat is very useful.

attached the pictures of my diamond running android 1. if you really want to try the images and have a place for other people to share please let me know and i'll upload the big image files as well as the initrd. :)

cheers !

- lt

[keerttttt]

Hi lavender, great news. You can post the files on my webhost if you want. I have a internet address with enough bandwith and speed.
PM Me.

[balsat]

1. diamond's built-in mmc cannot be partitioned (at this time). it is not because of linux btw. the problem is WM will not be able to recognize the partitioned mmc in diamond. - lt


Hmmm thats strange... I have partitioned my mmc in linux (fdisk), just remember to keep the first partition a vfat for Windows.
I got 2 partitions the first is a 2gb vfat and the second is a 2gb ext2, windows can read the vfat, and Linux read booth

[skiwithpete]

Great news Lavender!!!

[Willis111]

if you really want to try the images and have a place for other people to share please let me know and i'll upload the big image files as well as the initrd. :)

cheers !

- lt

I've got a bit of hosting space available. PM'd you the info.

[maejrep]

i was really excited when i saw cr2's formula. but when i looked closer i had the same question as netripper that there was no where to derive the real freq or divisor value from those parameters. also tried the formula for 144kHz parameters but i found the A4 value was different from what i got from the code. i thought 3 shift values could be wrong (missed by 1).
I used haret to dump a0 and a4 (on raph800), and I see:
HaRET(27)# pdump 0xa86000a0 0x08
a0 a4
a86000a0 | 0001fff4 fff50b44 | ....D...
{ ..., 0x20000 - 0xc , 0xfff50b44, ... }
Does this mean anything at all? I still don't know what the frequencies should be, but I noticed that those values weren't listed in your data structures.

[Untouchab1e]

Fantastic work Lavender. I run a web hosting service, so if you need a place to host your files or whatever, just let me know and Il set something up ;)

[utku]

Hi i tried this (http://www.netripper.com/raphael/20081204-01_raph_diam_android_v0.8.exe)

and it gave me the following error. maybe it means something to you

drivers/rtc/htcosys.c unable to open rtc device (rtc0)
trout_pwrsink_set:STUB!

[Untouchab1e]

Maybe he is just the type of guy who only speaks up when he feel like he got something to say.

Keep up the good work guy's!

Lavender is the real deal ;) Dont worry!

Anyhow, I was thinking.. This thread is about Linux and Android on the Diamond, and I feel its currently really messy considering people here are also working on porting Android to the Raphael. Maybe we should organize this a bit? Make a separate raphael thread or something?

Cheers!

[sblop]

Lavender is the real deal ;) Dont worry!



Im glad to hear :-)

[abc12345]

Lavender is the real deal ;) Dont worry!

Anyhow, I was thinking.. This thread is about Linux and Android on the Diamond, and I feel its currently really messy considering people here are also working on porting Android to the Raphael. Maybe we should organize this a bit? Make a separate raphael thread or something?

Cheers!

I was thinking that too

[Willis111]

Perhaps Lavender.t, like most other people who are actually do the dev work on this project are a little busy over the X-mas/New Years holidays with family and friends and we should be happy with whatever work they do regardless? Hopefully when he gets back from a little more of his vacation he'll upload his android 1.0 images and it will be a little easier to find out where we are and what hardware is working in the new version due to newer drivers being included. I think up until this point we've been rushing towards getting android 1.0 to boot for that reason, no need to re-write that's included in the actual operating system release since we've basically been running a beta up until we can get our grubby little hands on lavender.t's patches.

Edit: GJ Lavender.t

[lavender.t]

sorry for the delay folks. my internet went down last night just as i was checking my messages and trying to pack files up to upload.

thanks willis111 for lending his space. now you can download the 40MB .tar.gz from http://xda.willboost.com/. i myself have a microsoft office live host but it just allow me upload at most 10MB a time ... well, what can i say ... :(

this gz packags is by no means to make it easier to run in your phone (at this time). as i have said it is still in a very preliminary stage. very few things work. i just tried the calculator and it worked :) the key map doesnt work yet and the virtual keyboard always gets in my way which is kinda annoying ... no network ..., and no way to bring it back once the LCD is off after idle for a while ...

that being said, now the instructions: :)

1. unpack the tar.gz and copy the files to your phone. you will need haret.exe which is not included. make a dir named "tmp" from the root of the "internal storage". copy the 3 .ext2.img files into it.

2. now you might want to modify the default.txt a little bit. in my zImage i added 2 kernel parameters:

msmsdcc_id=2
msmts_calib=0x7a.0x5e.0x35a.0x37f

msmsdcc_id is the SDC number (1 - 4) of the "internal storage". my DIAM100 uses 2. and i think other diamond phones should use the same. so unlikely you'd need to change this.

msmts_calib is the touch screen calibration. i'm just lazy to click the tiny red dots every time booting up linux. so this parameter just allow me to do that once for all. i'm not sure if all the diamonds have the same calibration (but i think they are hopefully similar). so what you can do is to remove this parameter setting for the first boot. after calibration you will see the xmin = xxxx, ymin = xxxx .... values. these are hex numbers of your calibration. write them down. and next time, put in default.txt:

msmts_calib=<xmin>.<ymin>.<Xmax>.<Ymax>

note, they are dot (.) separated, NOT comma (,).

3. after booting up linux, you need to run /init.android to launch android.

well now i think you are pretty much ready to give it a try. one of the reasons for this package is that i hope it would be useful for other android 1.0 port since here we have a relatively clean cut of android and linux.

enjoy, and happy new year! :)

- lt

[lavender.t]

I used haret to dump a0 and a4 (on raph800), and I see: ...

thanks for the info maejrep. i'm getting the same from my diamond phone running haret.

HaRET(1)# pdump 0xa86000a0 0x20
a86000a0 | 0001fff4 fff50044 0001fffa fffb0044 | ....D.......D...
a86000b0 | 0005fcff fd04005c 0005fcff fd04005c | ....\.......\...

i guess it should be the same for yours too. A0/A4 are for SDC1, A8/AC for SDC2 ... and so on. i'm still yet to find out how to decrypt these numbers :(

from the logs you sent me last time, i'm worried that raph might be quite different from diam on this aspect. or possibly there are something different in the wince BSP of raph that touched some registers haret doesnt reset.

- lt

[fredfredburger123]

thanks lt - worked a treat

look forward to any more updates

am no where near an expert on this - just keen to see it running on the diamond

am always happy to test as have a spare one knocking about

thanks again!

[ronni.rasmussen]

Amazing work and thanks for the download works fine on the diamond with the deault.txt file.

Gonna be the phone geek at the party tomorrow thanks to you guys.

Happy new year yall!

[NoBody20013]

hi,

great work here!

can anybody tell about the performance? feels it faster than WM?


greetings

[ronni.rasmussen]

hi,

great work here!

can anybody tell about the performance? feels it faster than WM?


greetings

To early for me to say, cause no buttons is working for me so get stuck everywhere i got cause i cant go back. But what i've seen so far between the 10 soft resets i would say yes. But it would also be a HUGE suprise if it wasnt faster than WM :)

[xmoo]

msmsdcc_id is the SDC number (1 - 4) of the "internal storage". my DIAM100 uses 2. and i think other diamond phones should use the same. so unlikely you'd need to change this.


What about Xperia?? Storage Card

[fredfredburger123]

yeah if anyone can work out anything with returning to the homescreen that would be great! :)

[fredfredburger123]

is good :0)

[keerttttt]

I'm getting an error""
mmc0: new mmc card at address 0001
mmcblk0: mmc0:0001 000000 3932160KiB
mmcblk0: unknown partition table

And then it hangs there...

[ronni.rasmussen]

I'm getting an error""
mmc0: new mmc card at address 0001
mmcblk0: mmc0:0001 000000 3932160KiB
mmcblk0: unknown partition table

And then it hangs there...

Get the same message but it works fine so did you remember to:
make a dir named "tmp" from the root of the "internal storage". copy the 3 .ext2.img files into it. ???

Cause it seems like it has a problem with the ext2 files or the Intermal storage.

[Willis111]

msmsdcc_id is the SDC number (1 - 4) of the "internal storage". my DIAM100 uses 2. and i think other diamond phones should use the same. so unlikely you'd need to change this.


enjoy, and happy new year! :)

- lt
When I try and run your package I run into an error @

mmc0: new MMC card at address 0001
mmcblk0: mmc0:0001 ZLEODM 3932160KiB
mmcblk0: p1 ( p5 ) mount: mounting /dev/mmcblk0 on /mmc failed: Invalid Argument


followed by a bunch of errors about how it couldn't mount /mmc/tmp/system.ext2 etc etc

I can then calibrate the screen but of course can't init.android since it can't find the files.

Running on a telus cdma touch diamond


Edit: When I change the mmc address to 1 it crashes out after trying to set the mmc clock and about 3 trout messages w/no valid info

[keerttttt]

Get the same message but it works fine so did you remember to:
make a dir named "tmp" from the root of the "internal storage". copy the 3 .ext2.img files into it. ???

Cause it seems like it has a problem with the ext2 files or the Intermal storage.

stupid, must be that if i create a folder on my phone from my pc it capitalises it.
testing again now...
Edit: tested it, still no luck . same error... will check some more.

[keerttttt]

Allright, its working now.
I replaced the files using activesync instead of using it as disk drive.
That worked fine, and now it's showing me android 1.0.

Great job :D

Edit, works pretty fast also, smooth animations and everything. impressive!
Also great that i don't have to calibrate myself :P

Allright, another update:
it looks like android only boots up when i connect it with usb to my computer. the other times it hangs during the process and i don't get a busybox terminal.

[abc12345]

It works partially on my fuze, it stops at
mmc0: Host does not support reading read-only switch. assuming write-enable
mmc0 new SDHC card at address 41b7
mmc0: p1

[maejrep]

HaRET(1)# pdump 0xa86000a0 0x20
a86000a0 | 0001fff4 fff50044 0001fffa fffb0044 | ....D.......D...
a86000b0 | 0005fcff fd04005c 0005fcff fd04005c | ....\.......\...

i guess it should be the same for yours too. A0/A4 are for SDC1, A8/AC for SDC2 ... and so on. i'm still yet to find out how to decrypt these numbers :(

Hmm, not quite... Only a0-a7 are the same. Probably due to diam only having 2 real sdcc devices?

sd1
HaRET(27)# pdump 0xa86000a0 0x08
a0 a4
a86000a0 | 0001fff4 fff50b44 | ....D...

sd2
HaRET(29)# pdump 0xa86000a8 0x08
a86000a8 | 0019f3ff f418005c | ....\...

sd3
HaRET(30)# pdump 0xa86000b0 0x08
a86000b0 | 0001fffd fffe007c | ....|...

sd4
HaRET(31)# pdump 0xa86000b8 0x08
a86000b8 | 0019f3ff f418005c | ....\...

from the logs you sent me last time, i'm worried that raph might be quite different from diam on this aspect. or possibly there are something different in the wince BSP of raph that touched some registers haret doesnt reset.

- lt
Agreed, there seems to be a lot of differences between raph and diam, and even between raph100 and raph800 (can't wait to see what raph500 brings :))

[abc12345]

there is going to be a raph500?

[Arktronic]

RAPH500 is Verizon's Touch Pro.

Major differences are to be expected between GSM and CDMA Raphaels. After all, they have completely different radio stacks, so why not throw in a few more changes? ;) I'd expect the differences between 500 and 800 to be minimal, though, just like with 100 and 110.

[chlinkchlink]

Can anyone update the wiki with last linux for newbies like me ?

[Rusher]

Sprint HTC Diamond boots android 1.0 - OK. Colors messed up :(

[tdmusic]

Boots fine on my UK Diamond, awesome work Lavender (and everyone else who has contributed, of course!)

I get some weird graphical glitches - like random pixels on the screen? The first time, it took a while to boot (just on the "glowing robot" loading screen) and whenever I tapped the screen, a pixel would light. The second time, I noticed that when I pressed the # key in the dialer some random pixels would light?! Strange.

Also, the on-screen keyboard flickers in Android (I think this has already been mentioned).

We really do need a way to get back to the home screen too ;-) I remember reading that in 0.8, the stylus-in sensor functioned as a home button. This doesn't seem to be the case here - is there any way we can change the key mappings?

Anyway, far from usable, of course, but looks lovely and runs very smoothly in terms of loading, animation, etc. I'm really impressed.

I don't know much C, so I'm not sure how much I could contribute towards development, but my Diamond is a spare (company) phone so I'm willing to test whatever on it.

Awesome work guys, can't wait to see this get to the level it is on the Touch Pro or whichever handset it was :)

[xxxxchrisxxxx]

Don't know much (anything really!) about the technical side of things - just posting to say that I followed the instructions and this booted on my uk TMobile MDA Compact IV

Thank you :)

[Untouchab1e]

I dont reckon the files Lavender posted would work on the Touch Pro? But as Lavender said, its a pretty clean port of Linux and Android, so it Im hopefull that NetRipper or someone else can make it work on the Raph100 as well?

[tdmusic]

Yeah, I don't think I mean the Touch Pro... there was a Youtube vid of one HTC WinMo handset which seemed to be running Android pretty smoothly anyway!

Edit: I mean the Vogue

[lupine_85]

Nope, definitely doesn't work on the RAPH100 - same stopping-point as mentioned previously.

[pauldgroot]

Wiki updated by me.

Check it out here (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux).

[touch of jobo]

[...] as i have said it is still in a very preliminary stage. very few things work. i just tried the calculator and it worked :) the key map doesnt work yet and the virtual keyboard always gets in my way which is kinda annoying ... no network ..., and no way to bring it back once the LCD is off after idle for a while ...
enjoy, and happy new year! :)
- lt

I managed to bang together a single root fs booting into Android 1.0 from ram (..only moments before being able to access storage, so I guess it's no longer useful to pursue).

Results are pretty much the same.
- Screen times out, no way to bring it back. Per default, the display timeout is set to 1 minute. You can prevent this from happening in Settings -> Sound & Display -> Display Settings -> Screen Timeout -> Never Timeout.
- Hitting the 'end' key (on the vkeyb) has the same effect as letting the screen timeout: it dies. Any other key seems to go 'back' (..so I can't really type anywhere).
- The analogue clock doesn't run. The dgital clock in the status bar does. (I see that in the pics on the wiki (http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux#Progress), it does run..)
- The wallpapers seem to be too small for the screen: there is some black on the top and bottom, and when panned to the right also on the right side.
- In the bottom-left corner is a small overlay saying 'Safe mode'.
- When dragging the tab up or down, I would think that the lower half should be darkened (behind a dark semi transparent layer) but it isn't. The background only gets dark when the tab is completely up.

About the vkeyb getting in the way: In this kernel (http://members.home.nl/j0b0/zImage-jobo), the vkeyb is draggable. (..but it doesn't recognize the boot option for touchscreen calibration. A .diff for that (http://members.home.nl/j0b0/linux-htc-msm-2.6.25-jobo20081215.diff.txt) is linked to from this post (http://forum.xda-developers.com/showthread.php?p=3044529).)

With the latest clock-7x00.c on a raph100, I can now see the removable SD card, not the built-in storage. ..and mount and boot into any partition on it.
When I put the initrd on the SD card and boot from it, it doesn't get past the screen with the glowing android. (Sometimes it stops glowing after a while and it hangs completely, sometimes it keeps glowing forever.) I shall test with the fs images you posted and see if that makes a difference.

-Jobo

[kal1990]

hey guys,

i tested it on my diamond....it works fine....

but how come i back to the home screen when i do:Settings -> Sound & Display -> Display Settings -> Screen Timeout -> Never Timeout.
is there any way?

[plbelanger]

This is awesome work guys. Keep it up. I look forward to the day when everything functions correctly. Has anyone considered techy or social network sites to try and get the word out and get more people involved in development?

[LordKiwi]

the touch pro needs something a little different.
/dev/mmcblk0 points to the sd card.

for it to boot into the gui you need to copy the ext2.img file to /tmp map on the sd card.
And the init file in initrd.droid.cpio needs to be edited.


# mount the filesystems
mmcdev=/dev/mmcblk0
mmcdevp1=/dev/mmcblk0p1 #Add this line
system_img=/mmc/tmp/system.ext2.img
data_img=/mmc/tmp/userdata.ext2.img
cache_img=/mmc/tmp/cache.ext2.img

sleep 3
mknod ${mmcdev} b 179 0
mknod ${mmcdevp1} b 179 1 #Add this line too

mkdir -p /mmc
mount -t vfat ${mmcdevp1} /mmc

#changed device in the mount line from mmcdev to mcdevp1

mkdir -p /system
mkdir -p -m 0771 /data
mkdir -p -m 0770 /cache

this way the raphael boots android 1.0 from it's sd card.


Click here to download it. (http://members.upc.nl/m.hoek9/initrd.droid.cpio.gz)

[AstainHellbring]

the touch pro needs something a little different.
/dev/mmcblk0 points to the sd card.

for it to boot into the gui you need to copy the ext2.img file to /tmp map on the sd card.
And the init file in initrd.droid.cpio needs to be edited.




this way the raphael boots android 1.0 from it's sd card.

can you upload your edited initrd?

[sensorama]

This is awesome work guys. Keep it up. I look forward to the day when everything functions correctly. Has anyone considered techy or social network sites to try and get the word out and get more people involved in development?

i thougt this was a social network site for techy nerds. ;)

btw great work lav.

[Sl4sher]

the touch pro needs something a little different.
/dev/mmcblk0 points to the sd card.

for it to boot into the gui you need to copy the ext2.img file to /tmp map on the sd card.
And the init file in initrd.droid.cpio needs to be edited.




this way the raphael boots android 1.0 from it's sd card.


Click here to download it. (http://members.upc.nl/m.hoek9/initrd.droid.cpio.gz)

Thanks, that got it booting for me.
Unfortunately now that there's no screen calibration, Android seems to think the digitizer is on upside down so I can't type the command to start the gui. Is there any way to fix it?

EDIT: Actually now that I look at it, regardless of which initrd (original or kiwi's) I use it fails to mount the mmc using either device. I'm on a sprint raph 800. Maybe the sd card has a different name? Would it be possible to change the init file to mount the filesystems from internal storage (as in program memory, not the diamond's internal storage)?

[abc12345]

the touch pro needs something a little different.
/dev/mmcblk0 points to the sd card.

for it to boot into the gui you need to copy the ext2.img file to /tmp map on the sd card.
And the init file in initrd.droid.cpio needs to be edited.




this way the raphael boots android 1.0 from it's sd card.


Click here to download it. (http://members.upc.nl/m.hoek9/initrd.droid.cpio.gz)

It doesn't work for me, I get no such file or directory for all 3 imgs

[Vector-SS]

It doesn't work for me, I get no such file or directory for all 3 imgs


These should all be found under the imgs folder.

[Willis111]

Unfortunately now that there's no screen calibration, Android seems to think the digitizer is on upside down so I can't type the command to start the gui. Is there any way to fix it?


Remove the screen calibration from your default.txt and manually calibrate the screen @ boot.

[be_vigilant]

Wow looking good thanks for all your hard work guys.

Just came across this on youtube and seems to be the exact same version lavander created.

Is Lavander associated or are they claiming credit for his work?

http://uk.youtube.com/watch?v=VR-tsHQhPRo

edit...OK at the end they give thanks?......

[liamhere]

just watched the clip....strange
his favorite clips says it all really :eek:

[racedout]

Can confirm I just got it working on my FUZE from the sd card. Had to have it plugged in USB to get it to work though. The on-screen keyboard flickers but it is really cool booting 1.0!!! Keep up the good work and thanks!!!

[butterdori]

Yes he does mention lavender, but he also says 'powered by valvasoftware'... strange hmm....

[touch of jobo]

[...] And the init file in initrd.droid.cpio needs to be edited.
this way the raphael boots android 1.0 from it's sd card. [...]
Yes, that works for me (raph100). I had to adapt it a bit to mount an ext2 filesystem in stead of vfat, on the 2nd partition in stead of the 1st.
Things that are different:
* I get an error: "Sorry! Application Settings (in process com.android.settings) is not responding."
* Waaay more apps. :)
* Typing on the vkeyb now works and no longer has the side effect of back/cancel.
* Analogue clock runs.
Thanks,
-Jobo

[Willis111]

it seems I can't mount my mmc w/lavender.t's or touch of jobo's files. both drop me out to busy box saying mounting failed. when I ls /dev/mm* I see mmcblk0 listed, when I try and mount it manually to /mmc which also exists it says no such file or device.

Edit: Running a cdma telus touch diamond

[lavender.t]

Wow looking good thanks for all your hard work guys.

Just came across this on youtube and seems to be the exact same version lavander created.

Is Lavander associated or are they claiming credit for his work?

http://uk.youtube.com/watch?v=VR-tsHQhPRo

edit...OK at the end they give thanks?......

what the *!! i'm NO valvasoftware. i'm just myself. :mad:

and i wont take the credit for this either... many people, like netripper, druidu, dcordes, etc, had done lots of important work before i even start to read this board. i just came in at the right time to put what's been working so far together. and that's it ...

c'mon folks, if you find this useful, please *DO RESPECT* other people's hard work. for myself, i just did this for fun.

- lt

[dabs]

Hi all,

Sorry to clutter up the diam thread with yet more raph questions... Maybe at this point this thread should fork, as the differences between diam and raph seem to be becoming more of an issue...

With that said, here goes:

I am struggling to get this to boot on a RAPH100 (o2 UK variant).

I am using the initrd LordKiwi kindly posted and it boots as far as the busybox shell (only when usb is connected), but fails to mount the imgs.. I tried the original initrd and see the exact same issues..

eg:
...
mmc0: new SDHC card at address d555
mmcblk0: mmc:d555 SU06G (card size)KiB
mmcblk0: p1 <p5>
mount: mounting /dev/mmcblk0p1 on /mmc failed: Invalid argument

Then tries and fails to mount the three img's and (obviously) fails with No such file or directory errors then loads busybox.

Any ideas anyone? Quite probably I have made a stupid mistake somewhere, but cant see where. I dont need an "idiots guide" and am prepared to get my hands dirty messing with the initrd (or partitioning the SD card to add a second ext2 partition), but if anyone knows what going on it would save me a lot of time fumbling with the initrd and then waiting for the damn thing to soft-reset ;)

@jobo,

What are the pros/cons of using vfat imgs vs. an ext2 partition?

Could you post the procedure you followed (or even better zip/gz up the whole lot and up it someplace, but I know thats a lot to ask) for running from an ext2 partition.

I assume you extract the imgs to their relevent folders on the root of the ext2 partition then hack the initrd to mount everything where it should be, but a little tut would be very nice.

btw. Major props to all those involved in the project so far!

Thanks in advance :D


ps. Whats the deal with it crashing immediately before loading busybox when usb is not connected? Odd...

[lavender.t]

Yes, that works for me (raph100). -Jobo

wow ! first confirmation of the sd working in raph ... cool and thanks jobo !

yes it is meant to be the external sd for raphael. from netripper's info raph and diam uses (mostly) the same board design -- even the sd part. the difference is that diam built the sd inside.

i was wondering why the raph phones had no luck getting the clock right ... great to hear that at least one model works. :)

- lt

[Willis111]

Any ideas anyone? Quite probably I have made a stupid mistake somewhere, but cant see where. I dont need an "idiots guide"

I think I could use an idiots guide. The device is always created but for some reason it's never actually there and can't be mounted.
I'm running w/setup like

/device
/device/internal storage/tmp/haret files
/device/internal storage/tmp/images

(just dumped everything all in the same dir same as ss's on the wiki)

in async mode and sd card mode both flop out after trying to mount the sd card/images and I'm not able to manually mount the drive, therefore images. I've tried jobo's image and it does the exact same thing. Has anyone got this running on a cdma diamond and can post a step by step/are all you guys that have gotten this running using gsm versions of the phone?

[lavender.t]

Hi all, I am struggling to get this to boot on a RAPH100 (o2 UK variant). I am using the initrd LordKiwi kindly posted and it boots as far as the busybox shell (only when usb is connected), but fails to mount the imgs..

hello dabs, for the external sd it really depends on how the sd is partitioned and formatted. so other people's sds might be different from yours base on how they are used.

if you can plug the sd into a linux box it would be easier to see what's in there. sorry that i didnt build fdisk into the busybox so it would be a little tougher to do in the phone.

if the linux box recognizes your sd card, e.g., as device /dev/sdd, then you can do a "fdisk -l /dev/sdd" to see all the partitions and types. if it doesnt show any, then the entire card is used as a single storage media (which is the case for my DIAM110). in this case, running "dd if=/dev/sdd count=1 | hd" will dump the first sector and this may give some hints what type of file system is formatted in there. but if you havent done anything special to your card, by default it should be vfat.

regarding the type of file systems to use, android natively uses yaffs2. the reason i used ext2 is that the file system needs to support symbolic links (i guess without symlink it might work as well, but that's a waste of space and i dont have much time to try), which is a no for vfat. and i had trouble getting yaffs2 over loopback working in my kernel image. the downside of ext2 loopback over vfat is the performance -- disk access has to go thru an extra layer. but the good thing is you dont have to repartition/reformat your card for a different file system so you dont lose any existing data.

hope this helps,

- lt

[lavender.t]

Has anyone got this running on a cdma diamond and can post a step by step/are all you guys that have gotten this running using gsm versions of the phone?

hi willis111, have you tried what i sent you in the private message ? we can chat by PM or in the irc. dont wish to turn this thread into support ... ;)

- lt